Self-Regulation in Emerging and Innovative Industries

I. Introduction

The concept of industry self-regulation is nearly as old as industry itself. We know, for instance, that early guilds of merchants and craftsmen set and enforced quality and trading standards to maintain the reputation and integrity of their goods.^[1] And despite the more recent expansion of governmental regulation, self-regulation remains a core legal and regulatory mechanism in our society.^[2] However, existing theoretical accounts of self-regulation are incomplete. In an under-recognized phenomenon, self-regulation has proven critical in shaping the trajectory of emerging industries—that is, in the kinds of innovative industries that have in recent years significantly disrupted and transformed our world economy.^[3]

In this Article, we study several cases of emerging transformative industries—the nascent hydrogen fuel industry, the historical development of the internet, the early days of hydraulic fracturing for oil and gas, and the currently burgeoning artificial intelligence field—to shed light on self-regulation’s functions in the context of especially innovative, technical, and sometimes urgently needed emergent industries. Many of these industries tend to be networked, requiring multiple new pieces to fall into place simultaneously to take off. We examine how self-regulation, in concert with some degree of public regulation, can contribute to innovation and growth in these industries, particularly by (1) striking a practical balance between stabilizing certainty and creative uncertainty; (2) addressing unregulated externalities and filling jurisdictional voids in global industries; (3) encouraging competition for self-regulatory primacy; and (4) helping scattered components of emerging industries coordinate around shared goals. Viewed through the lens of an industrial policy that favors innovation and economic transformation to address societal challenges and imperatives,^[4] self-regulation has an important role to play that extends and even transcends its paradigmatic function as a risk-mitigation tool.

Our analysis is unique in bringing together two literatures—the self-regulation and innovation literatures—that are usually siloed from each other.^[5] There is a robust literature exploring the conditions under which effective self-regulation is likely to work as a partial or complete substitute for government regulation,^[6] and there is likewise much work specifying the appropriate relationship between self-regulatory standards and public “meta-regulation.”^[7] There is also a growing literature on the value of self-regulation in the environmental and energy contexts, where informational and political challenges sometimes prevent more direct government regulation.^[8] Another literature explores the possible value of regulatory uncertainty within nascent technological areas, but primarily with a focus on public regulation and private law, not on self-regulation.^[9] And a large innovation literature analyzes the drivers of nascent industries but tends to give short shrift—if any attention—to how regulation (whether public or private) can complement these drivers.^[10] All of these strands of the literature, though important in their own right, overlook the central role that self-regulation can play in enabling the emergence of innovative industries at the boundaries of our changing economy.

Our analysis of these case studies also shows that self-regulation’s innovation-fostering functions are not a given, but instead need to be cultivated. While self-regulation frequently played the leading role in shepherding the industries we study from infancy to maturity, public regulators often operated in the background to steer the emerging self-regulatory ecosystem in public-serving directions. For instance, public regulators may spur self-regulatory standard setting, help address conflicts among competing industry standards, and fill gaps in standards when necessary to provide a backstop for public benefit.^[11] Our analysis finds that this symbiotic relationship between self-regulation and public regulation is especially essential for networked industries with multiple components that require coordination—in other words, the dominant type of emerging industry in today’s networked economy.^[12]

This Article unfolds as follows. Part II of the Article recounts some of the relevant history of U.S. self-regulatory regimes. Part III presents a starting theoretical framework of the conditions that tend to support self-regulation’s innovation-fostering functions. Part IV is the core of the Article: a case-study driven investigation of how private standard setting and self-regulation is developing (in the case of hydrogen and large language models (LLMs)) or has developed (in the case of other areas, such as the internet and hydraulic fracturing). We ask and answer questions about how well the theory outlined in Part III matches the reality: that is, how the various standard-setting efforts have been coordinated (or not), the effectiveness of standards at resolving the primary risks and technical issues presented by emerging industries, and what can be done to spur a dynamic relationship between industry self-regulation and innovation through public and private standards. Part V then offers a synthesis and insights for those thinking about the role of self-regulation in other emerging industries not studied here.

II. The Role of Self-Regulation in the United States

Much regulation in the United States involves a government entity issuing prescriptive standards and imposing consequences on private firms or individuals. Regulation is in fact a much larger concept, however, covering all “pressures and policies” administered by both public and private actors.^[13] One deviation from the paradigmatic case of command-and-control regulation is self-regulation, which involves any regulatory target imposing “commands and consequences upon itself” either individually or collectively.^[14]

A. Self-Regulation’s Ubiquity

Self-regulation is, in practice, extremely diverse, but it is commonly conducted through standard-setting organizations (SSOs), such as the American National Standards Institute (ANSI), the American Society for Testing and Materials (ASTM), and the American Society for Mechanical Engineers (ASME).^[15] Increasingly, SSOs like the International Organization for Standardization (ISO) play an important role in harmonizing business practices across international boundaries.^[16] Self-Regulatory Organizations (SROs)—organizations formed by trade associations or groups of firms to both write and enforce standards for those firms—are similarly common.^[17] Often, self-regulation is a part of a more layered regulatory ecosystem involving rule-makers, rule-takers, and in-between entities that might be characterized broadly as “regulatory intermediaries.”^[18]

Whatever specific form it takes, self-regulation is pervasive, providing critical guidance in a wide range of sectors, including, for example, aviation, consumer products, cannabis, construction, and manufacturing.^[19] Indeed, self-regulation has a long and storied history in the United States, and it is poised to become even more important in the coming decades as the need for regulation, particularly in critical, highly technical areas with coordination challenges, outpaces government’s ability to provide it.^[20]

In an example of a foundational industry heavily reliant upon self-regulation, one of the more storied (and studied) loci for self-regulation is the financial sector.^[21] Historically, much of the securities industry has been regulated by two familiar private organizations—the New York Stock Exchange (NYSE) and the (private) Financial Industry Regulatory Authority (FINRA)—that are overseen by the U.S. Securities and Exchange Commission (SEC).^[22] Emerging business models in the financial sector have so far been handled—to the extent that they have been handled at all—via self-regulation. For instance, FinTech—the practice of using cutting-edge technologies in financial services^[23]—relies heavily on self-regulation to “increase investor protection and enhance the reputation of the industry.”^[24] The creation of a SRO for tradable digital assets, such as cryptocurrencies, has been advocated for by many for its potential to provide guidance and bolster the industry’s reputation amidst destabilizing scandals like the collapse of the cryptocurrency exchange FTX.^[25]

Also worth noting is the substantial amount of self-regulation in the environmental regulatory space, which complements relatively well-developed and comprehensive public governance.^[26] For instance, chemical manufacturers, under the leadership of the Chemical Manufacturers Association, helped institute a program called “Responsible Care,” which created “practice codes” to help firms set voluntary performance standards for hazardous chemicals management.^[27] A variety of non-governmental organizations have also helped to spearhead industry self-regulation in sustainable forestry and fishing practices.^[28] And what Vandenbergh and Gilligan refer to as “private environmental governance,” such as eco-labeling, is an example of self-regulation.^[29]

The energy sector, too, has long relied on industry self-regulation, perhaps in large part due to the technical complexity of many issues that arise in the industry. Most famously, nuclear power producers turned to self-regulation in the wake of the Three Mile Island near-disaster in 1979, creating the Institute of Nuclear Power Operations (INPO) to manage a performance regulation regime for nuclear safety.^[30] Less well known are the private “grid regulators” that ensure that electric power is reliable.^[31] Additionally, private Regional Transmission Organizations (RTOs) govern electric transmission systems and manage markets for wholesale energy on a regional basis, also with (deferential) Federal Energy Regulatory Commission (FERC) oversight.^[32] Beyond electricity, in the oil and gas production and transportation context, American Petroleum Institute (API) standards—more than 800 of them—govern most aspects of the industry.^[33] Federal and state regulators have incorporated many of these standards directly into public governance.^[34]

While most of these examples are old news, self-regulation appears to be poised to play a role in several new venues. The emerging hydrogen-fuel sector, for instance, is a critical sector to decarbonizing heavy industries such as freight transportation and aviation, and one that we also investigate in this Article.^[35] Hydrogen will likely be no exception to the history of self-regulation in the energy sector. Hydrogen is a fuel produced from natural gas, coal, plants, or water that can be produced through zero-carbon processes and used for industrial, commercial, and residential purposes.^[36] Despite these opportunities, already, major gaps are apparent in efforts to apply existing government regulatory programs to hydrogen. These programs were designed to address problems in other energy subsectors, meaning that hydrogen will require improved approaches.^[37] And given the unique nature of hydrogen (it ignites more easily than natural gas or gasoline, for example), the emerging hydrogen industry cannot simply repurpose the many private standards that exist for other subsectors of the energy industry.^[38]

Another somewhat nascent but rapidly evolving industry investigated here—artificial intelligence (AI), and particularly open large language model (LLM) AI such as ChatGPT—is primarily self-regulated through an SRO called the Frontier Model Forum, formally launched in July 2023 by the major OpenAI firms.^[39] This Forum researches AI safety, identifies “best practices and standards,” and is designed to support “information sharing among policymakers and industry.”^[40] While discussions continue about a more public regulatory response to the challenges posed by the rapid growth of AI and LLMs,^[41] for the time being, most regulation will likely emanate from SROs like the Frontier Model Forum. As with hydrogen, AI poses new and different risks from somewhat similar technologies in its areas (such as simpler computing applications) and requires a substantially new set of standards. And it is a global industry with no jurisdictional boundaries—the same issue posed by the internet, which emerged primarily within a self-regulatory system, as we explore here.^[42]

We could go on, but the basic point is clear by now: self-regulation plays a major role in the U.S. economy, and indeed in the world economy. As part of a core suite of “flexible” regulatory tools,^[43] self-regulation has been a key component of a broader societal turn away from state-centered regulation and towards what some have called “responsive regulation”^[44] and “collaborative [regulatory] governance.”^[45] These approaches include regulatory action, norm-building, and collaboration by a variety of non-state actors.^[46] Much of the impetus for self-regulation undoubtedly arose from an impulse to deregulate heavily regulated industries, or at least to dilute the state’s control overregulation. This is certainly the case with much energy-system regulation, which shifted only in recent decades from a heavily regulated public utility model to the more de-centered, self-regulatory model we see today.^[47] However, as the AI and FinTech examples show, it is often the case that self-regulation is the only game in town; it fills major gaps left in existing government regulatory regimes or reinvents these regimes as new business models emerge and attempt to disrupt or displace existing sectors.^[48] These disparate examples of self-regulation across industry sectors suggest that self-regulation serves many purposes in our modern economy, far beyond simply relieving regulatory burdens for firms. In Parts III and IV we argue that a major function of self-regulation, whether intended or not, is the fostering of the conditions that enable innovation in emerging industries. But before doing so, we turn to the functions of self-regulation explored in existing literature.

B. Existing Literature on Self-Regulation

Self-regulation emerges for a variety of reasons, and its effectiveness depends on an equally varied array of factors, as the literature has explored in depth (typically for established industries). The literature links drivers of self-regulation to pressures such as “non-legal interventions like social movement activism,”^[49] a desire to boost the reputation of an industry,^[50] attempted avoidance of potential legal liabilities or the threat of more heavy-handed government regulation,^[51] an informational advantage for self-regulators over government agents,^[52] or a combination of these forces.^[53] In addition, self-regulation might be viewed as a desirable arrangement by government regulators, who may be aware that they lack the necessary expertise or resources to run a successful regulatory program themselves,^[54] or who may see self-regulation as a pragmatic style of iterative, flexible regulatory governance.^[55] If self-regulation works—i.e., if the standards address risks and externalities efficiently and effectively and are enforced—they can offer many benefits. Proponents of self-regulation often tout the “responsiveness” and “reflexivity” of self-regulation as an alternative to traditional models of regulation.^[56] Finally, particularly when it is paired with standardization, self-regulation can provide a stable platform for innovation, although, as we will discuss, this rationale for self-regulation is perhaps the most undertheorized.^[57]

Despite its potential benefits, self-regulation does not always succeed. Studies of the chemical industry’s Responsible Care program—an environmental self-regulatory regime—suggest that it “largely failed, at least in its early years and at least when measured in terms of reducing members’ toxic emissions.”^[58] Indeed, a more recent study of Responsible Care found that plants that participated in the program emitted 15.9% more toxic pollution than similar non-participating plants.^[59] Naturally, the question arises: why does self-regulation sometimes work and sometimes not work? On one account, firms develop and join self-regulatory frameworks for their symbolic value and often lack economic incentives to actually change behavior or “raise the bar” above a perfect convergence between self-regulation and managerial objectives.^[60] Against this conclusion, some research finds that self-regulation can form an institutional structure that generates enforceable norms, among other effective regulatory alternatives.^[61] These more informal pressures can, in theory, motivate firms to engage in meaningful self-regulation. However, the empirical literature on self-regulation suggests that only “the pressure of deterrence” generates effective self-regulation.^[62]

While these firm-level studies focus largely on incentives and institutions, it is also the case that characteristics of the regulated industry can greatly impact self-regulatory governance independent of these other variables. One strand of the literature examines the firm attributes that must be present to incentivize self-enforcement by firms, observing that self-regulation is likely to be more effective with smaller, more tightly knit industries that can easily police bad apple firms who might wish to free-ride on the reputational benefits of private standards without actually complying with them.^[63] Again, the Responsible Care program is instructive. Several studies conclude that one reason Responsible Care was less successful than other comparable programs, such as INPO, is that it attempted to coordinate self-regulation across a highly heterogeneous group of firms, some of which had incentives to free-ride on the program’s reputational benefits while not actually making meaningful changes to practices.^[64] By contrast, INPO involved a “much smaller and more homogeneous” set of firms, which gave “individual firms . . . more of a common interest” and gave industry leaders more power to “rein in potential outlier firms.”^[65] As with any collective action problem, the greater degree to which individual actors have divergent preferences, the more difficult the achievement of a coordinated self-regulatory strategy is likely to be.^[66] It is still possible for intra-firm self-regulation to take its place, but these programs are more apt to be scattershot and to fail to drive meaningful behavioral change due to the lack of group pressure.

The deficiencies of self-regulation cannot always be solved internally—say, by tightening lines of communication or connection in an effort to address collective action problems and free-riding. In these cases, the addition of public governance through meta-regulation is often needed.^[67] A large literature explores the art of balancing private and public involvement in standard-setting. Saule Omarova categorizes a variety of forms of public involvement in the self-regulatory project, such as “sanctioned” self-regulation, in which government actors must approve private rules; “mandated” self-regulation in which government requires self-regulation; and “co-regulation,” in which “public agencies and private market actors cooperate in the creation, implementation, and enforcement of rules.”^[68] In the financial context, James Fanto observes that some types of meta-regulation—such as a proposed public Financial Service Oversight Council assigned to project and prevent risks in the financial industry—might simply reinforce a particular form of top-heavy institution that is unable to accurately predict and respond to risks, whether it is private or public.^[69]

In summary, there are numerous types of self-regulation and rich accounts of the factors that may drive self-regulation and make it worthwhile, as well as accounts analyzing when self-regulation is likely to be “successful” along a variety of metrics, such as efficient risk reduction. But there is less discussion of the role of self-regulation in upstart industries—those that require research, innovation, experimentation, and, if successful, extensive scaling up. We take up this task in the following part.

III. Self-Regulation in Emerging and Innovative Industries

Much of the existing literature on self-regulation reviewed in Section II.B is cast at a general level. The question typically addressed within this literature is when self-regulation will occur and when it will be successful across a wide range of firms. Due to this focus on generalizable theories, the literature so far has missed some of the distinct questions that arise around self-regulation specifically in emerging industries, where there is a premium on innovation and flexibility, as well as potential network and coordination benefits of self-regulation for scaling up operations.

There is a push for self-regulation in specific emerging industries such as the peer-to-peer sharing economy^[70] and in FinTech “regulatory sandboxes”^[71] where space for innovation is highly prized. Yet these movements are only beginning to generate evidence in support of a broadly applicable theory of self-regulation in emerging industries. There is a need for more careful thinking about the specific opportunities and challenges of self-regulation within this space. Simply put, self-regulation is likely to play a very different role in industries that are poised to grow substantially than in industries that resort to self-regulation to solve a reputational or legal risk problem. This moves us beyond questions of the effectiveness of self-regulation as a risk-mitigation tool that can meaningfully complement more traditional forms of public regulation (i.e., whether self-regulation serves as mere window dressing or a more genuine deterrent to risky industry behavior). We extend self-regulation analysis to fundamental issues involving the circumstances under which self-regulation is likely to play a fundamental role in supporting industrial development while effectively addressing risks and public concerns associated with such development.

This part first reviews relevant literature to build a theory of when and how self-regulation will prove important to innovative, emerging industries. It then turns to the parallel question of how public regulation and self-regulation can play complementary roles in fulfilling these functions of self-regulation.

A. The Distinctive Opportunities and Challenges of Self-Regulation in Emerging Industries

For an emerging industry, it can be difficult if not impossible to spur government institutions to be proactive about supplying needed regulation. The hydrogen industry is a case in point: despite consensus-based calls for more federal regulatory action to fill gaps in existing regulations created by new hydrogen technologies and practices,^[72] the federal government has shown limited initiative in taking on even relatively uncontroversial regulatory updates. The combination of scarce resources for regulatory housekeeping and the fact that the hydrogen industry is still in its infancy make it highly unlikely that the federal government will prioritize regulation in this area, especially if doing so means forgoing other needed regulations in more established industries. Innovation-focused entities such as the National Renewable Energy Laboratory, Argonne National Laboratory, and Sandia National Laboratories have highlighted regulatory progress and gaps, summarized existing standards and regulations, and directly supported efforts to write gap-filling regulations.^[73] Yet agencies such as the FERC and Pipeline and Hazardous Materials Safety Administration (PHMSA)—which control core regulatory frameworks in other, more established subsectors of the energy economy—have made few, if any, meaningful changes to adapt existing frameworks to hydrogen.^[74]

Even if the federal government were to provide comprehensive public regulations for the hydrogen industry and other upstart industries, it is not necessarily clear that this kind of regulatory regime would be preferable to a more organic convergence on coordinated industry standards. Self-regulation, with its built-in focus on efficient, responsive, flexible, comprehensive, and coordinated regulation, may be better suited for industries that are in rapid stages of evolution and learning. Self-regulatory standards can, in other words, strike a delicate balance between regulatory certainty and flexibility, a particularly important feature for emerging industries in which regulatory learning and updating are key.^[75] Likewise, self-regulation might present greater opportunities for more bottom-up competition over standards than a more top-down public governance approach would allow. Furthermore, for networked emerging industries that produce complex coordination challenges, private actors—particularly SSOs that take a birds-eye view on all facets of an industry, from production through distribution, and often with informal guidance from public regulators—tend to produce the type of coordinated standards that help to guide all components of an industry forward at the same time. For example, ASTM International, in promoting its comprehensive private cannabis standards, boasts a “360-degree approach” to regulating and supporting the expanding industry.^[76]

This is not to say that public standards should play no role in emerging industries, however. Just as meta-regulation is sometimes needed to address the collective action or free-riding problems associated with industry self-governance, government involvement can address these problems and even spur innovation in the context of emerging industries.^[77] But it is to say that self-regulation may be uniquely well-suited to emerging industries, and that the advantages it creates in this domain are worthy of greater attention. In what follows, we highlight and categorize structural advantages like these that can make self-regulation particularly attractive at the early stages of a needed industry’s development, and we explore the complementary role of public governance.

1. Competition in Standard Setting.

One question that is decidedly present in the context of emerging industries is competition for self-regulatory primacy. This trend stands in stark contrast with many established industries, which often already have trade associations or other networks that can supply top-down standards that cover the entire industry.^[78] With emerging industries, it will often be the case that there are many potential private standards. In some cases, there are too many standards, and this drives the formation of an SSO or SRO, or public intervention. Take the example of standards for fitting fire hoses to fire hydrants—of which there were more than 600 in 1904, leading to collaboration between the National Fire Protection Association and U.S. National Institute of Standards and Technology to form a national standard.^[79] Even when too many standards generate problems—such as the Great Baltimore Fire of 1904, in which hoses did not fit to fire hydrants—the top-down organization that addresses these problems and creates a uniform standard has a diverse range of examples to draw from in this case, potentially generating a more effective and efficient standard.^[80]

Standards-based competition—as with public regulatory competition in federalist or decentralized governance systems^[81]—could generate more effective, efficient controls that best address the diverse needs of regulated entities, and perhaps even more so than in the public context. Indeed, for better or worse, competitive self-regulation sometimes fills a space that would otherwise be occupied by states competing to attract industry or residents through beneficial regulatory diversity.^[82] The literature on interjurisdictional regulatory competition (public or private) suggests that although there may sometimes be a “race to the bottom” in terms of regulatory stringency and efficacy, this is highly context dependent.^[83] In some cases, interjurisdictional regulatory competition, by creating a market for “location rights,” leads to a “race to the top.”^[84] That is, if people demand quality regulation and are able to vote with their feet by relocating to a different jurisdiction that better matches their preferences, theory suggests we should end up with regulation that is welfare enhancing.^[85]

There are good reasons to think that the benefits of regulatory competition are more likely to accrue in the context of competition for self-regulatory primacy than in the classic case of interjurisdictional governmental competition. To the extent that incentives for self-regulation exist at all, choice among potential SSOs and SROs presents a far more realistic version of “foot-voting” for rules than does Tiebout’s public governance example, in which residents shop by moving to packages of local government goods, services, and taxes that best meet their needs.^[86] One of Tiebout’s core assumptions was that voters are mobile and can express preferences through exit and entry. But voters are often tied to jobs, family, mortgages, or other localized obligations that make moving quite difficult.^[87] Self-regulatory standards, unlike public rules promulgated by municipalities, remain unburdened by artificial jurisdictional lines. Private standard-setting bodies and SROs can and often do offer their standards to any relevant subscriber or qualifying member—namely, any industrial actor on the globe that fits within the technical confines of the industry subject to the standards. This makes locational mobility of the regulated actor a non-issue.^[88]

Beyond mobility, Tiebout’s assumption that foot voters are knowledgeable about the different governance options may be truer in the self-regulatory setting, where actors are relatively sophisticated and might actively research and shop for private standards in an effort to appease potential investors and consumers, or, perhaps, avoid threats of overly stringent public governance. In fact, the limited empirical evidence that does exist about competition over self-regulatory governance, drawn from the self-regulation of the Internet of Things field, suggests that potential subscribers to competing SSOs are savvy shoppers. Potential subscribers, the data show, choose SSOs that provide a “broad membership base” and “low costs and fast development” of standards.^[89]

Finally, while those building from Tiebout’s work tend to assume (often incorrectly) that public governments meaningfully and methodically experiment with different policies,^[90] private standard setting organizations often do engage in more purposeful experimentation with standards and assessment of their efficacy, in part to advertise results to firms that might subscribe to the standards.^[91] Indeed, as we explore in Part IV, many members of the hydrogen industry are developing their own standards and associated technical experiments designed to test and demonstrate the efficacy of various standards.

To be sure, there are some assumptions that may remain unmet in standards-based competition. For example, Tiebout assumed that no economies or diseconomies were shared among communities, whereas local governments’ regulations regularly have positive and negative spillover effects.^[92] The same is true for private standard setting. Although standards can be proprietary and more easily limited to subscribers, their benefits—such as creating overall investor confidence in a subset of actors within an emerging economy—cannot be easily cabined.^[93] Furthermore, some broader assumptions of regulatory experimentation remain unrealistic in both the private and public contexts, including, for example, the questionable assumption that those designing and enforcing a governing regime will honestly measure and share negative results in addition to positive ones.^[94] Nevertheless, there are reasons to believe that competition for self-regulation in emerging industries is a feature, not a bug, and one that would necessarily be forfeited if public regulation were to displace standards too quickly.

2. Striking a Critical Balance Between Regulatory Flexibility and Certainty: Substance and Jurisdiction.

Industry actors and investors frequently cite the importance of regulatory certainty or predictability, arguing that without clear, relatively durable guiding standards it is difficult to attract adequate investment or make large capital outlays. Industry actors do not want to invest in capital and procedures for compliance only to find these investments obsolete within several months or years. For emerging industries in which jurisdictional authority is often unclear (federal versus state control, for example), clarity as to who will regulate also plays an important role.^[95]

Some degree of regulatory certainty is therefore particularly critical for emerging industries, which need large initial investments to undertake the expensive and risky endeavor of building a network of infrastructure from scratch.^[96] Yet as scholars in the energy sphere have observed, there are many types of regulatory certainty, and in fact some degree of uncertainty or vagueness can create flexibility. Flexibility is also essential for emerging industries, particularly those at an early technological development stage, to allow them to experiment.^[97] As such, the special situation of emerging industries like hydrogen or AI creates a need for a “Goldilocks” approach to the balance between certainty and uncertainty. Here again, self-regulation can be better suited to find this balance.

a. Substantive Clarity and Risk Mitigation. Regulatory uncertainty often means that there simply are not many (or any) regulations on the books guiding a particular industry—thus making it unclear whether the industry will be permitted or banned or how and whether it will be regulated. This poses problems for the public—in the form of potentially unmitigated risks—and for industry, in the form of uncertainty and concerns about a high-profile negative event that will stifle further development of the industry.^[98]

Regulatory uncertainty arises in several different ways, sometimes through a lack of any regulation—posing a question of when and to what extent an emerging industry will be regulated, if at all—and most commonly through the limited application of existing regulations to a new industry. For hydraulic fracturing, for example, one environmental group persuaded the federal Eleventh Circuit to hold that the Safe Drinking Water Act—which requires regulatory approval prior to the injection of substances underground—applied to hydraulic fracturing (Congress later exempted hydraulic fracturing from the Act).^[99] For AI, there are large grey areas of regulation, with some existing public standards applying in only a limited way (such as FTC rules).^[100]

Regulatory uncertainty of all types involves a climate in which new rules might emerge once an industry is developed, presenting a potentially drawn-out legislative or regulatory process and the risk (to industry) of stringent, costly regulation in the future. For the public, there are concerns that the regulation will be inadequate to address risk, or, if there is high demand for the emerging product, that it will stifle innovation.^[101]

For this type of substantive uncertainty—the question of whether the industry will be regulated at all, or banned, and how it will be regulated—self-regulatory standards can play an important role. These standards can fill in gaps and produce a somewhat predictable path forward for the industry. For the many types of regulations in which omissions are not a ban, the lack of any public regulation threatens to produce a regulatory “Wild West” scenario, in which industry proceeds with no regulation. This, in turn, can produce negative effects and understandably alarm residents near or those affected by the emerging industry (not to mention risk-conscious investors) and can trigger moratoria and bans. A clear set of self-regulatory standards that are consistently enforced can alleviate a regulatory void, address emerging risks, and potentially avoid the drastic public remedy of a ban. Indeed, particularly in the early stages of an industry, the industry actors themselves likely have superior knowledge of risks as compared to officials.^[102] And if public regulation eventually emerges, it often incorporates some of the self-regulatory standards already in place, thereby avoiding dramatic substantive changes.^[103]

This very progression occurred within the oil and gas industry, for which there were few state-specific environmental regulations of the industry before a new type of hydraulic fracturing dramatically expanded in the mid-2000s, causing production to boom.^[104] A number of local governments with nonexistent or sparse oil and gas ordinances, and certainly no comprehensive regulations specific to hydraulic fracturing, reacted with moratoria and bans.^[105] Others allowed the development to proceed and eventually incorporated a substantial number of the API standards that industry already followed, such as regulation of the safe construction of wells to prevent leakage of pollution underground; standards for design and operation of waste impoundments; and technologies that prevent dangerous well explosions during drilling and hydraulic fracturing.^[106] With many of the outright bans and moratoria facing trouble in court, the private API standards approach quickly became the standard.^[107] But without the standards in place, the regulatory vacuum that might have existed would likely have changed the dynamics of how courts and other policymakers viewed heavy-handed, precautionary regulation.

As we will discuss in more detail in Part IV, a great deal of substantive regulatory gap-filling of this sort already appears to be occurring in the hydrogen context. The industry recognized the lack of a clear standard for the safety of hydrogen pipelines as an impediment to growth as early as 2007, when ASME was in the process of developing consensus standards for such pipelines. As the president of ASME’s technical branch noted in 2009, although the “traditional approach to standardization has included writing prescriptive standards only after technology is fully established and after commercialization has been completed,” a “lack of standards in a particular area can actually create a barrier to commercialization.”^[108]

Similar concerns abound for the AI industry, which is very aware of the risks posed by its technology and the harm that could ensue (in addition to, potentially, more heavy-handed regulation than the industry would prefer). Before launching an SRO in July 2023 to address these risks, the Chief Executive Officer of OpenAI testified before Congress in May 2023, acknowledging that the technology could do “significant harm to the world” and supporting government oversight.^[109]

b. Stable Jurisdictional Control. Beyond substantive omissions or a lack of clarity, another type of regulatory uncertainty is jurisdictional, involving the serious issue of which level or levels of government will regulate and whether an industry will be subject to conflicting regulatory regimes at different levels.^[110] This type of uncertainty can in some cases be beneficial; e.g., by allowing the emerging industry to experiment in a variety of locations without a national standard that has incorrectly predicted the risks. This jurisdictional openness can support the same type of experimentation with standards discussed previously, which can again produce effective, responsive, and efficient rules.^[111] But jurisdictional uncertainty, in particular, as opposed to a mere substantive diversity of standards as an industry emerges, can also pose a formidable obstacle. The emerging cannabis industry is a case in point. As Professor Jonathan Adler observes, this policy area is a key example of competitive federalism, in which states compete to offer packages of policies preferred by voters. Yet federal law, which impedes everything from bank lending to tax credits in the area of legalized marijuana, often gets in the way of what could be productive competition.^[112] Indeed, ASTM International plays a key role in providing a comprehensive set of private cannabis standards to support quality assurance, marketing, testing, and other aspects of the industry within a complex federal jurisdictional climate.^[113]

The internet represented—and AI now represents—an even more uncertain jurisdictional extreme, presenting a form of industry that transcended sovereign boundaries and defied traditional jurisdictional control.^[114] It is hard to imagine any industry that has not been touched by internet technologies, or that will be touched by AI in the future, yet there are no sovereign actors with comprehensive authority over all economic activities that might rely on these technologies.^[115] Here, self-regulation of a broad suite of practices and technologies, complemented by public standards, was and is essential because it can respond to issues in the use of these technologies wherever and whenever they occur, as we discuss in Part IV.

c. Flexibility. Although emerging industries need reasonably predictable rules and knowledge of the jurisdictional level at which these rules will be promulgated and implemented, flexible rules that allow for learning and updating are also critical. Here, there is a delicate balance between certainty and durability, on the one hand, and the need to be able to modify rules when there is a good reason to do so, such as when new risks become apparent or when new technologies come on the scene and take the industry in an unpredicted direction. Substantive flexibility is built into self-regulatory regimes, with most standard-setting organizations having regular schedules for reviewing and updating standards. As shown in Part IV below, the SSOs with hydrogen standards have in some cases already amended or updated these standards several times, or even canceled them in light of new data on superior self-regulatory approaches.^[116]

In the oil and gas context, the American Petroleum Institute has three models for updating standards: (1) “periodic maintenance,” in which standards must be reviewed “when technological changes affect their currency or at least once every 5 years”; (2) “continuous maintenance,” for which revisions may be proposed at any time; and (3) “stabilized maintenance,” under which standards for mature technologies and practices must be reviewed every ten years.^[117]

Periodic amending and updating of standards produces calculated flexibility—a form of regular, scheduled modification of standards that does not unduly compromise certainty. Public regulation can and sometimes does achieve a similar result through requirements for the periodic sunsetting and review of rules,^[118] but empirically speaking, such rules appear far less prevalent than built-in review in the self-regulatory context.

Self-regulation also offers important jurisdictional certainty and flexibility—giving a national or international SSO or SRO primary control while providing room for state, federal, or international government actors to enter the regulatory sphere, either by displacing some standards or, as often occurs, incorporating many of them by reference.

B. Simultaneous Production of Standards to Address Coordination Challenges

Beyond facing regulatory uncertainty, emerging industries often present distinct coordination challenges. These challenges arise even if the industry is not exceedingly complex from a supply chain perspective. Take the example of a widget that has one ingredient and one end use. Producers and transporters of the ingredient; the widget manufacturer; and transporters, distributors, and end users of the widget still must emerge, somewhat simultaneously. The simultaneous development of these distinct industry components is itself a challenging proposition calling for some coordinating feature, be it an organized group of industry actors, a vertically-integrated firm, or government-led coordination through an economic development organization or similar entity.^[119] But spurring separate and disparate governmental agencies with control over different parts of the supply chain to regulate at the same time—either independently or in a coordinated fashion—is similarly challenging.

Bottom-up self-regulation is the more likely approach to the development of coordinating standards in various components of an emerging industry. Larry Lessig describes a “coordinating standard” as “a rule that facilitates an activity that otherwise would not exist” and limits liberty simply to ensure that the activity can exist.^[120] Lessig’s simple example of a coordinating standard is one that requires driving on the right-hand side of the road.^[121] Nascent or once-nascent industries such as hydrogen and the internet create more complex examples, in which numerous standards must converge to allow vehicle users to locate hydrogen fueling stations with compatible fueling technologies, for example, or internet users to connect to a network and understand the letters that identify a webpage.^[122] Lessig observes that coordinating standards can be top-down or bottom-up.^[123] With the internet and, so far with hydrogen, the standards have been very much a bottom-up effort, with industry leading the way toward coordination, albeit with government actors sometimes spurring or at least aiding such efforts.^[124]

This is not to say that governments always fail in coordinating regulation for emerging industries or economic trends. But comprehensive sets of rules that cover multiple components of an industry and enable the industry simply to exist in its essential networked form, from manufacturing to transportation and distribution, appear to be far more common in the private context.^[125] This is likely because industry associations that cover multiple facets of the supply chain—or even disparate associations—have strong incentives to coordinate to write a comprehensive set of standards, as occurred in the context of oil and gas and its extensive set of API standards.^[126] Alternatively, industry may work with an independent SSO to gain a bird’s-eye view, as has occurred for cannabis and ASTM International^[127] and appears to be the case for hydrogen so far.

Industries with interconnected or networked risks seem particularly motivated to coordinate in the standard-setting context, as evidenced by the electricity industry. The North American Electric Reliability Council (NERC) is composed of electric utility members, and these members formed NERC in the 1960s to address the cascading risks that permeate the electrical system.^[128] The electric grid—generation, transmission, and distribution—is tightly interconnected, and electricity supply depends on a delicately balanced flow of electricity through wires at all times. Any instability at any physical point in this vast interconnected network can accordingly cause large regions to lose power for days at a time, leading to billions of dollars in economic loss. NERC subsequently developed a relatively comprehensive set of standards for the reliability of all grid-connected infrastructure, from transmission lines and substations to power plants and distribution wires, which endured for decades without public oversight.^[129]

Self-regulation alone, however, might not achieve the degree of standards coordination necessary for a complex supply chain in need of innovation within all links that form the chain. Here, governments play an important, top-down complementary role in producing or encouraging the formation of comprehensive, more coordinated standards (public or private or both) that address the many facets of an emerging industry, particularly when they view this industry as critical to public goals. As we discuss in Part IV, the Department of Energy (DOE) is following this path in the hydrogen context, with an explicit mission of “coordinating and accelerating the efforts of major standards and model code development organizations and regulatory agencies so the required standards, codes, and regulations for hydrogen technologies can be prepared and adopted to facilitate commercial applications of these technologies in a timely manner.”^[130]

Beyond spurring the development of coordinated standards (public or private or both), governments often work to clarify and educate new industrial players on the existing and emerging standards that apply to them, thus attempting to ease regulatory burden and hasten the development of the industry. In the context of the U.S. hydraulic fracturing boom, some states took both of these approaches, promulgating relatively comprehensive regulatory changes and regulatory menus that clarified the suite of regulations that applied to drilling and hydraulic fracturing.^[131] This was due to a combination of citizen concerns about adequate compliance with regulation, the fact that many industrial actors were rushing in from other states and were unfamiliar with state-specific rules, and the desire of some states to encourage economically beneficial oil and gas development by clarifying the rules. In the hydrogen context, as discussed in Part IV, the federal government has created permitting checklists and guidance for industry participants while supporting efforts to fill in standards-based gaps.

Finally, some governments are so motivated to spur industrial transition that they also take measures to implement and ensure compliance with the private or public regulations, or both, on behalf of industry, bringing governmental assistance with coordination of standards to the highest level. This is one form of what Coglianese & Mendelson call “meta-regulation.”^[132]

---

In summary, self-regulation seems particularly well poised to address some of the critical challenges faced by emerging industries that are technically complex and face steep coordination challenges. Self-regulation, perhaps even more so than public standards, can produce effective, responsive standards through regulatory competition. Self-regulation can provide certainty in the face of regulatory gaps, yet needed flexibility through periodic updating as an emerging industry learns of new or different risks. And although governments can continue to play an important role in coordinating public and private standards as they emerge, industry is even more motivated to play this part and is perhaps more likely to be centrally engaged in organizing and integrating the standards that address numerous components of a complicated supply chain. Self-regulation, in other words, can be a key enabling component of a nascent industry’s efforts to grapple with coordination problems.

Despite these benefits, there are important limitations to self-regulation that are important to note in the context of emerging industries. Extensive reliance on self-regulation without adequate attention to self-regulatory decision-making structures can give individual firms too much power in the process, thus risking problems of capture.^[133] This could be particularly problematic during the early stages of an industry, when even those most familiar with the risks still have much to learn about safe design and deployment of technologies. A classic example of inefficient technological lock-in is the QWERTY configuration of most keyboards, which, according to some accounts, was adopted because of the preferences of a now obsolete profession of telegraph operators. We are still beholden to their preferences to this day.^[134] Furthermore, even well-intentioned, highly-informed industry actors can have blind spots to important regulatory areas, areas that might be better detected by more neutral government actors with a bird’s-eye view. Yet self-regulation, complemented by public governance and well-designed SSO standards development processes, seems to hold great promise for emerging industries such as hydrogen.

This discussion gives short shrift to the many other factors supporting innovation, in part due to space limitations yet also in part purposefully. The innovation literature extensively explores the role of intellectual property, traditional property rights, regulatory exemptions, and, more recently, the role of commons and peer production in the emergence of an industry.^[135] But it says too little about self-regulation as a potential driver of innovation.

IV. Self-Regulation of Nascent Industries and Services

The theory that self-regulation can help to spur and support the growth of an emerging industry, as suggested in Part III, is best tested by examining how self-regulation plays out on the ground in emerging industries. This part takes up that task, assessing the role of self-regulatory standards in a series of case studies of emerging economies—specifically, hydrogen fuel, artificial intelligence, and the internet—zeroing in on whether these case studies accord with the theory in Part III.

This part focuses on hydrogen and the internet because these sectors are, or once were, critical emerging industries in which self-regulation was or is pervasive. The internet transformed the global economy, just as hydrogen fuel could be a game changer in the global energy system. The opportunities were and are obvious. These innovations also pose or posed risks: in the case of hydrogen, physical safety risks and economic risks, and in the case of the internet, social risks pertaining to speech and democracy.^[136] This part also briefly explores AI because it, too, is emergent and critical, in that the innovative technology underlying AI could offer extensive societal benefits, yet also poses grave threats.^[137] Most of these economies also happen to be networked, meaning that multiple components must simultaneously emerge for the economies to be successful. While we focus on the innovation and scaling up aspects of these economies in the self-regulation context writ large, we also explore the role of self-regulation in supporting the networking component of these industries.

As detailed below, our case studies, while by no means reducible to cookie-cutter versions of the theory laid out in Part III, do provide substantial support for many of the theoretical benefits of self-regulation identified there. Every sector has its own story to tell, but there do seem to be commonalities that suggest that self-regulation can be useful for emerging industries at the frontiers of our economy, and that this self-regulation can work best when government itself does not remain entirely on the sidelines, but rather coordinates the work of self-regulatory actors.

A. Hydrogen as a Textbook Case of Growth Through Self-Regulatory Support

The development of industry standards in the hydrogen space largely reflects a pattern of standards within similar emerging industries over the past two centuries: an individual or company pioneers a new technology or practice; a nascent but growing industry endeavors to persuade the public of the need for this innovation and of its safety and/or utility; and somewhat coordinated industry standards, alongside or prior to governmental regulations, emerge. Industry writes these standards (or relies on an SSO to write these standards) for several purposes: to persuade the public of the safety of the innovation; to avoid the potential for an industry-wide collapse if a high-profile safety incident were to occur; and to produce somewhat coordinated standards to make implementation of the innovation feasible.

1. Why a Focus on Hydrogen?

The hydrogen economy—one that has faltered in the past and is now receiving billions of dollars in global support, including in the bipartisan Infrastructure Investment and Jobs Act (IIJA) and Inflation Reduction Act (IRA)^[138]—serves as an ideal case study of the role of self-regulation in complex emerging economies. In 2023, hydrogen was an industry that barely existed but that needed to grow substantially and rapidly to meet public climate governance goals.^[139] Indeed, hydrogen figures largely in most strategies for decarbonization of the economy—particularly for so-called hard-to-decarbonize activities, such as heavy industry, air travel, and shipping, which cannot be easily electrified.^[140] In every scenario identified by the Deep Decarbonization Pathways Project, so-called “green” hydrogen (hydrogen manufactured through electrolysis using renewable energy power) use will have to expand massively, providing low-carbon liquid fuel for transportation and gas fuel for building heating and industrial activities.^[141] Increasingly, proponents are also touting how hydrogen could revolutionize the storage of energy, a challenge that pervades the energy sphere.^[142] The DOE’s comprehensive “Hydrogen Program Plan” estimates that there is likely to be a $2.5 trillion global market for hydrogen by 2050, with the U.S. share of the market amounting to $750 billion per year.^[143]

At present, the hydrogen industry is essentially undeveloped. While in the United States hydrogen is currently shipped through about 1,600 miles of pipeline and used in certain refinery and chemical manufacturing processes,^[144] many of the expected applications of hydrogen are entirely hypothetical at this point, and generation, storage, transportation, and distribution would all have to be scaled up drastically, and simultaneously, to meet the expectations of the Hydrogen Program Plan. Due to the central role of hydrogen in the energy transition, governments are racing to enlarge the hydrogen sector, investing approximately $76 billion in concrete global government funding for the industry, including billions in recent U.S. funding.^[145] The DOE recently announced the location of several “Clean Hydrogen Hubs,” which are to serve as demonstration projects for “accelerat[ing] the large-scale production and use of clean hydrogen.”^[146]

Despite the current enthusiasm and government support for hydrogen, a mountain of hurdles stands between these ambitions and the realization of a hydrogen economy. Chief among them is that hydrogen fuel is literally starting from the ground up, but will rather quickly have to become a networked, infrastructure-heavy industry. The development of supportive networks poses classic coordination problems,^[147] with industry players needing to collectively build new transportation networks (or access old ones, including retired natural gas pipelines) to grow beyond the initial hubs. Production is unlikely to emerge without committed end-use outlets and pipelines, and distribution infrastructure to access those outlets; end users, in turn, will not commit to purchase hydrogen without an assurance of steady availability and transport options for the product. The hydrogen economy is in some sense a “lumpy social good” that may not be supplied without careful coordination of the various complementary (and, in some cases, nonexistent) components of the industry.^[148]

Beyond coordination challenges, the hydrogen industry is somewhat path dependent: once a specific design for metals for hydrogen tanks and pipelines is selected, it will be difficult to turn back. Such path dependency is often cited as a factor militating in favor of self-regulation,^[149] and it will be exceedingly important to get the substantive standards right in the hydrogen space from early on to create certainty for the nascent industry, yet also preserve essential flexibility as more is learned. And finally, governmental jurisdiction over the many components of the hydrogen economy is far from clear,^[150] leading to regulatory gaps. Standards have the potential to promote cross-jurisdictional consistency while public entities iron out who among multiple layers of government will ultimately regulate particular aspects of the hydrogen economy.

Compounding coordination and path dependency challenges are environmental and human health risks. Many components of the hydrogen economy will require attention to safety concerns. Hydrogen is safer than other common fuels with respect to toxicity and some other metrics, but many public concerns remain.^[151] Transporting hydrogen through pipelines is also challenging because hydrogen embrittles the metals typically used for pipelines.^[152] For workers handling and storing hydrogen, suffocation is a very real risk.^[153] For many of these challenges, the unique characteristics of hydrogen will prevent simple adoption of existing regulatory approaches. Rather, rules and standards will have to be developed from scratch while researchers are continuing to update learning about hazards, and industry is experimenting with solutions. Federal and state regulations, moreover, have lagged well behind innovation in this space. Even relatively simple questions, such as whether and how to repurpose natural gas pipelines to facilitate transportation of hydrogen, are currently in flux and are unlikely to be addressed in government regulations any time soon.^[154] Instead, much of the regulatory governance of the nascent hydrogen economy will have to be provided through self-regulation, and, as we discuss below, much of that self-regulation is already taking place.

2. Hydrogen Self-Regulation Emerges from Early Fire Codes.

One of the most coordinated standards for hydrogen—and one that covers many aspects of the hydrogen supply chain—is housed within one of the oldest and most-used SSOs in the United States: the National Fire Protection Association (NFPA).^[155] Numerous states and local governments directly incorporate NFPA standards into their building, fire, and electrical codes.^[156] And the early history of NFPA standards for sprinklers and electrical systems was in some sense a test run for NFPA’s own eventual development of comprehensive hydrogen safety standards.^[157]

As explained by NFPA’s former Executive Director, a group of pioneering individuals formed the NFPA as individuals were first inventing and attempting to prove technologies for spraying water on fire—the predecessors to modern sprinkler systems—and the safety of electricity and electricity hook-ups. As new sprinkler systems took root in the 1870s, a number of different designs and standards emerged, as exemplified by nine different standards for the size of water pipes and spacing of sprinklers within buildings in the vicinity of Boston in 1895. This created a “plumber’s nightmare” and endangered the success of the fast-growing industry.^[158]

At the same time, five different U.S. standards addressed the “safe use of electrical equipment,” creating five different sets of rules for electricians installing such equipment and generating “confusion” and controversy.^[159] And the safety of electrical hook-ups was by no means proven or widely accepted by the public at this point; indeed, the 1893 World’s Fair included a “Palace of Electricity” that insurers were reluctant to cover prompting the organizers to hire an expert (William Henry Merrill) to prove the safety of the electrical equipment and hook-ups.^[160] Electrical manufacturers subsequently wrote to Merrill for certifications of their products as safe, spurring him to create a safety testing laboratory (ultimately to be named “Underwriters Laboratories”) supported by fire insurers and electrical manufacturers.^[161] But other private electrical codes were emerging at the same time, primarily written and administered by fire insurance underwriters and associations of electrical equipment manufacturers, thus contributing to the confusion noted above.^[162]

The late 1890s were therefore a scene of conflicting industry standards creating a confusing array of rules for electricians installing wiring and outlets and for plumbers installing sprinkler systems, all relatively new and innovative technologies at the time. Yet the confusion did not only arise on the installation side. Manufacturers were centrally concerned about easing installation so as to prove and sell their products, and fire insurance companies wanted more guidance and input on the preferable standards. Insurance underwriters and manufacturers accordingly gathered in two important convenings in 1896: one in which a National Electric Code was formed, and another in which members developed the NFPA.^[163] By 1897, the NFPA was established as the centralized private body for all standards relating to fire protection—electrical, sprinkler-related, and otherwise.^[164]

NFPA’s focus on fire codes made it, more recently, a natural locus for standards development for hydrogen technologies, which presents serious fire risks due to its high flammability. To prove commercially viable, the nascent hydrogen industry will have to persuade the public and insurers of the safety of its products, particularly given historic high-profile incidents such as the Hindenburg explosion.^[165] Although further study has revealed that the explosion was primarily caused by the flammable chemical coating on the zeppelin, the public often focuses on the hydrogen that fueled the flying machine.^[166] Fast forward to the twenty-first century, NFPA responded to this need for public reassurance by promulgating, over the course of almost a decade, one of the most comprehensive hydrogen safety codes in existence in the world—a code structured to be the national code for hydrogen safety in the United States.^[167]

The Hydrogen Technologies Code (or NFPA 2, as it is referred to) provides “fundamental safeguards for the generation, installation, storage, piping, use, and handling of hydrogen in compressed gas (GH2) form or cryogenic liquid (LH2) form.”^[168] Unlike some SSOs, NFPA makes its standards, including NFPA 2, available for free online reading, although it retains copyright and restricts printing, allowing it to continue to generate revenue from “the thousands of users who pay for their copies of the code.”^[169] It addresses many common problems in hydrogen applications, such as setback requirements from fuel storage tanks.^[170] Originally appearing in 2011,^[171] NFPA 2 has been updated several times, including in 2023, and was incorporated by reference in the California Fire Code in 2014.^[172] As we discuss later, in recent years NFPA has been rounded out by a plethora of other hydrogen-specific private standards on more specialized problems in the emerging hydrogen economy.^[173]

Unfortunately, this bloom of self-regulatory activity happened a few years too late to allow it to help it realize an early opportunity to support the emerging hydrogen industry. Prior to this recent push for hydrogen technologies standards, one pioneering U.S. state attempted to jumpstart one specific component of the hydrogen economy—vehicles powered by hydrogen fuel cells, which in turn required hydrogen fueling stations. This effort, which occurred without the support of many public or private regulations, such as NFPA 2 or the California Fire Code’s incorporation of hydrogen standards, was largely a failure. It produced lessons about why the more recent, second generation of more comprehensive hydrogen standards in the wake of NFPA 2 is likely critical to the attempted resurgence of a previously failed industry.^[174] We now turn to what might be learned from this cautionary tale about the risks to emerging industry from the lack of self-regulatory standard setting.

3. California’s “Hydrogen Highway”: A Cautionary Tale?

In 2005, Governor Arnold Schwarzenegger released a plan to scale up the use of hydrogen fuel cell vehicles in the state under California’s Zero Emission Vehicles regulation.^[175] While this Hydrogen Highway plan focused on just one of the many potential uses of hydrogen, it is still possible to glean broader lessons from this experiment regarding the importance of self-regulation in the emerging hydrogen economy. Indeed, this case may well be a cautionary tale about trying to scale an innovative technology without well-developed self-regulatory or public governance frameworks in place.

The California Hydrogen Highway Blueprint was intended to proceed in several stages, combining public and private initiatives to build a network of hydrogen fueling stations and assist automobile manufacturers in developing models of cars with fuel cells. The first phase set a relatively meager target of 50–100 fueling stations and 1,200 fuel cell vehicles by 2010. The second phase greatly scaled up the ambition, calling for 250 hydrogen fueling stations and 10,000 predominantly fuel cell vehicles; phase three called for a doubling of the vehicle count to 20,000.^[176] Years later, it is clear that almost none of this came about: hydrogen fuel cell vehicles are now a “legacy zombie technology” linked in the popular imagination to a “bad bet by the state.”^[177] The state has already spent $125 million on the program and fallen short of expectations, with just about 50 fueling stations and about 9,000 hydrogen-powered vehicles.^[178]

While some of the failures of the Hydrogen Highway can be attributed to the economics of the technology and the chicken–egg problem of building fueling infrastructure before there was consumer demand, at least some of the failure is likely also attributable to the infancy of the movement to develop standards at the time of the program’s rollout. Although the Blueprint document emphasized the need to jumpstart the process of developing codes and standards, this work was slow to catch on. It was not until 2014 that California adopted NFPA 2 in the California Fire Code and California Building Code (although, to be fair, NFPA 2 itself was not finalized until 2011).^[179] Several of the most important standards for fuel cell vehicles—such as Society of Automotive Engineers J2719, on hydrogen fuel quality for fuel cell vehicles, and J2760, on pressure terminology used in fuel cells—did not come out until after 2011 as well. What this means, practically, is that the few fueling stations that did exist in the early stages of the Hydrogen Highway project had to pass through permitting processes by reinventing the wheel for each project rather than simply following standard industry practices that would guarantee success, a process that surely added cost and uncertainty to efforts to build stations. Moreover, because of this diversity of standards, it was possible for different stations to operate using different technologies and specifications, surely a challenge for workers constructing these stations and a headache for would-be hydrogen fuel cell vehicle owners who reportedly run into many problems with out-of-service stations.^[180] This lack of standardization is analogous to the challenges faced early in the electricity and fire protection/prevention industries, when competing standards caused confusion for plumbers and electricians and failed to assure consumers of the efficacy and safety of what were then relatively new products.

Whether the more developed codes and standards that exist today could reverse the fortunes of the California Hydrogen Highway program is an open question. In 2020, the California Energy Commission authorized $169 million in new funds to open up 111 public hydrogen fueling stations.^[181] Now, developers have a wealth of materials to assist them in streamlining projects, including NREL-funded YouTube videos, providing essential information on relevant codes and standards and the permitting process.^[182] Still, it might be too late for the hydrogen automobile industry. Automakers have scaled back production of fuel cell vehicles in favor of battery-electric models, which seem to be taking hold as the dominant technology for light-duty vehicles.^[183] Perhaps this trajectory would have been different had the standards been developed at the outset. Indeed, it is possible that with more standards in place now and the Clean Vehicle Credit in the Inflation Reduction Act incentivizing new purchases, hydrogen vehicles may still have a future in the United States.^[184] That is especially the case because, as we demonstrate in the next subsection, the movement to produce standards is no longer in its infancy.

4. The Second Generation of Hydrogen Standards: Competition, Coordination, and Certainty in the Midst of Regulatory Gaps.

The current global push for hydrogen is far more ambitious than California’s hydrogen vehicle efforts. Countries around the world are investing billions of dollars in an entire hydrogen economy, envisioning the production, transport, and distribution of hydrogen to support a variety of energy intensive end uses in the residential, commercial, industrial, and energy storage sectors.^[185] Happily, an impressive suite of industry standards has emerged in the past decade (compare a compilation from 2011^[186] to one contemplated just seven years later in 2018^[187]), suggesting potentially more hope for this round of attempted innovation. But much progress remains to be made, as explored here.

In the United States, the effort to develop industry standards is increasingly a coordinated effort amongst SSOs. Members of the hydrogen industry from all portions of the supply chain have joined the Fuel Cell & Hydrogen Energy Association (FCHEA), with a mission of “[p]roviding the industry a voice in shaping regulations, codes, and standards to enable commercial growth, while ensuring the highest levels of consumer safety and satisfaction” among other purposes.^[188] FCHEA works to draft some standards—particularly in the hydrogen sector—but is primarily concerned with educating private- and public-standard setting bodies about hydrogen developments and providing a centralized portal of all of the global hydrogen standards currently in place.^[189] One organization has begun to create a singular set of comprehensive guidelines and information—a document rising to “platinum” status in an industry that involves so many coordination challenges.^[190] This coordination of standards may in part arise from the physical interconnectedness of the emergent hydrogen economy, requiring production facilities to be connected to storage, pipelines, distribution lines, and end users.

Interestingly, the government has also stepped up to assist these coordinated standards development processes while still leaving the matter to SSOs. Most notably, the DOE’s Hydrogen & Fuel Cell Technologies Office, along with several national laboratories (e.g., the National Renewable Energy Laboratories and Sandia National Laboratories), have worked closely with SSOs to improve and round out hydrogen codes such as NFPA 2.^[191] These efforts endeavor to “identify which regulators and agencies need to be engaged by . . . stakeholders . . . for future development of hydrogen technologies” and “identify the limits of federal oversight, in particular what types of systems would fall under state/local jurisdiction rather than federal.”^[192]

The evolution of industry standards for hydrogen alongside the slow and sometimes halting growth of the industry itself provides important insights into the role of self-regulation in supporting or limiting an industry’s trajectory, as explored below.

a. Competition and Coordination in Hydrogen Standard Setting. A well-populated set of standards has already emerged for an almost wholly theoretical hydrogen economy and is impressively varied—applying to most segments of the supply chain, ranging from production and storage to the safety of hydrogen in a variety of end uses, such as vehicles.^[193] There are 393 standards published globally, and 15 under development in 2022, suggesting a relatively robust competition in standard setting.^[194] This is further supported by the variety of organizations offering such standards, including competing substantive standards within the same regulatory space—NFPA is no longer the only cop on the beat. Organizations that have published significant hydrogen standards include the American Society for Testing Materials (e.g., fuel quality assurance), American Society of Mechanical Engineers (e.g., hydrogen tanks), Compressed Gas Association (CGA) (e.g., pressure vessels for portable metal hydride systems), CSA Group (e.g., hydrogen fueling stations), and International Organization for Standardization (e.g., fuel specification for PEMFC vehicles), and many other organizations have produced smaller sets of standards within this space.^[195] The National Renewable Energy Laboratory concluded as early as 2015 that private standards and government regulations “address all key aspects of system design, construction, operation, and maintenance” in the context of hydrogen fueling stations for vehicles.^[196]

Experts at Sandia National Laboratories, reviewing private standards and public regulations that apply to the broader hydrogen economy, identify only seven areas—out of dozens—that are not “hydrogen ready” due to a lack of standards, although some of these areas are quite consequential. For example, experts surveying the standards and regulations note that there are no regulations authorizing hydrogen import or export terminals; no regulations addressing the sale or distribution of hydrogen to residences; and no testing requirements for furnaces, boilers, and similar residential heating systems that could run on hydrogen.^[197] Similarly, there is a need for more technical data as remaining gaps are filled. This appears to be happening organically. For instance, ASME, which has played an important role in the formation of consensus standards, created a separate non-profit LLC to test the technical aspects of hydrogen, noting that “[a] series of technical reports addressing hydrogen infrastructure applications directly resulted in new ASME code rules for pressure vessels, piping, and pipelines specific to hydrogen.”^[198] These rules are particularly important as industry players contemplate how to retrofit existing natural gas pipelines so that they are appropriate for the transportation of hydrogen.

As the remaining gaps become smaller, we are seeing a turn to intra-standard competition. That is, private standards-setting organizations are even addressing flaws that they see in other standards. For example, the CGA officially expressed concerns that some NFPA rules for the setback of hydrogen storage tanks from nearby infrastructure were “unclear.”^[199] CGA accordingly published a Position Statement that it asked hydrogen storage tank installers to follow and asked the NFPA to clarify “unclear separation distances” between hydrogen storage tanks and other infrastructure.^[200] NFPA did amend its separation distance requirements in 2016, although it is not clear whether this was in direct response to CGA’s request. Rather, as early as 2010 NFPA had begun to reconsider the required distances between hydrogen storage tanks and other infrastructure and humans to base its standards more clearly on risk factors, such as the type and probability of hydrogen exposure rather than solely the total volume of hydrogen stored in the tank. NFPA built upon government (Sandia Lab) developed hydrogen models and updated the standards in 2010 and again in 2016.^[201]

In short, while there are still holes to be filled, the self-regulatory framework for hydrogen is quite robust, and certainly much more so than in the 2000s with California’s failed “Hydrogen Highway” experiment. Most major issues now have at least some degree of standardization from at least some SSO, resulting in a veritable quilt of self-regulation. While this state of affairs is preferable to ten years ago, it also points the way to new challenges of coordination and competition, and the balance between the two. Self-regulation in the hydrogen economy appears to be poised to turn to a period of refinement, competition, and experimentation to improve the existing suite of standards.

b. Simultaneous Production and Coordination of Standards Within a Vast Hydrogen Supply Chain: The Case for Active Coordination. Although letting one thousand flowers bloom in the standard setting context has produced relatively complete standards, new challenges are now emerging. A patchwork of competing standards can be confusing and expensive to comply with, among many other challenges. The “black start” of a highly networked economy from ground zero—requiring the simultaneous development of production, transportation infrastructure, and end uses—is a situation that desperately calls for the development of a comprehensive set of standards at multiple points of the supply chain. In the context of hydrogen, producers, storage companies, pipeline and distribution line builders, owners, operators, and end users all need minimum safety standards to begin operating, and few parts of the industry will function very effectively if the other parts are not up and running.^[202]

In short, there does seem to be a pressing need for more top-down coordination of hydrogen’s well-populated self-regulatory ecosystem. Interestingly, some of the law firms most active in the hydrogen space assume that government will have to produce a relatively comprehensive set of standards, arguing that “the federal government will need to incorporate hydrogen into its broader regulatory scheme for hydrogen to truly become part of the energy infrastructure in the U.S.”^[203] And indeed, government is playing a critical, albeit merely facilitative, role in the development of a coordinated U.S. hydrogen economy by working hand-in-hand with SSOs. An abundance of relatively uncoordinated standards, paired with some standards-based gaps, is a particular threat for hydrogen, given its potentially expansive use in the residential setting for cooking, heating, and backup generation. As the DOE notes, there are over 44,000 local government jurisdictions in the United States, all of which have their own building codes.^[204] Many states provide a uniform building code (comprising primarily private standards incorporated by reference) that local governments adopt, but local governments are not always required to follow this standardized approach.^[205] The DOE is working with private standards-development organizations to attempt to coordinate the plethora of disparate building code standards that could potentially apply to residential hydrogen uses.^[206]

As part of its effort to speed up and help coordinate the development of comprehensive, navigable standards for hydrogen, the DOE is also working to identify “gaps in the standards development process and provide methods to close the gaps,” “[p]roviding support for key international standards meetings,” supporting research and development efforts that provide the data needed to develop effective standards, and “[s]upporting the codes and standards adoption process,” among other coordinating activities.^[207] The DOE and its subunits are also endeavoring to ease industry compliance with a complex set of regulations that apply to numerous aspects of the supply chain, producing videos and literature on how to permit a hydrogen fueling station; charts that map out public and private hydrogen standards; checklists and permitting guides; trainings of code officials who review compliance with hydrogen requirements; and regular trainings on emerging hydrogen technologies and hub opportunities, the content of hydrogen standards, and status of newly developed standards.^[208]

Despite concerted governmental efforts in coordinating hydrogen standards, the bulk of the efforts appear to be emerging from industry associations, as we predict would occur in Part III. For example, the DOE refers viewers to an association-created database of hydrogen standards, which characterizes standards by the organizations that created them and the sectors to which they apply.^[209] This organization—the Fuel Cell & Hydrogen Energy Association—also supports industry standard development efforts. And as explored above, organizations such as NFPA and the CGA are actively working to develop more comprehensive standards and to address gaps in standards.^[210] Additionally, ASME formed a separate organization to specifically test the technical aspects of the industries for which it develops standards; it has used ASME Standards Technology LLC to test hydrogen risks and produce technical papers to directly inform the ASME hydrogen standards process.^[211] The precise balance between SSO- or SRO-led coordination and government-led coordination is being determined in real time, but the overall need for some degree of coordination is clear and is being met.

c. Evidence of Balanced Certainty and Flexibility Within Hydrogen Standards. A critical feature of regulations guiding all industries—but particularly those in the development phases—is the creation of regulatory guidance that is relatively stable yet also offers flexibility through periodic updating or standards. Stable standards provide some degree of certainty and predictability as industry invests in the capital and services needed to comply with regulation. Periodic updating of standards—not so frequent as to upset or overturn substantial investments in compliance, yet frequent enough to address new information about risks and regulatory needs—in turn allows for changes that may enhance consumer confidence and address industry concerns about overly costly or inadequately effective regulations.

The relatively comprehensive set of U.S. industry standards for hydrogen appears to be playing a critical role in filling some large regulatory gaps that produce uncertainties for a nascent hydrogen economy—particularly PHMSA’s failure to write regulations for the safety of pipelines that carry hydrogen and FERC’s lack of hydrogen pipeline construction standards.^[212] And industry associations are actively updating these standards as they learn of new risks, as demonstrated by NFPA’s updating of its hydrogen standards in several different years based on new government risk models.^[213]

B. Self-Regulation and Public Governance in the Growth of the Internet

A variety of industrial sectors beyond the energy sector offer similarly important insights into the potential for self-regulation to support emerging industries—particularly technologically complex industries with network effects.^[214] The internet is perhaps the most apt example. Here, too, was a largely undeveloped public good—a small yet critical information network pioneered by the military and, at first, offering extremely limited access by private individuals. As the internet grew, it represented even stronger network effects than hydrogen; it increased in value with every new connection of an individual to the literal information network, and it required a physical, interconnected infrastructural web of wires and hardware. Even more so than hydrogen, the internet was also a good with an extreme lack of jurisdictional clarity—indeed, a good that defied all traditional political boundaries. And as with hydrogen, the internet involved detailed technical questions that those closest to its formation were most familiar with.

The internet began as a government-led U.S. military endeavor that connected private government contractors with military computers.^[215] But as private individuals pushed for and obtained access to it, government involvement in the design and control of the internet substantially changed, while self-regulation expanded. The predominant, early assumption of experts was that this technologically complex, sprawling network required self-regulation. Commentators noted the need for flexibility of standards as technologies rapidly evolved and emphasized the highly technical nature of the internet, which called for private expertise in standards development.^[216] To a large degree, governments agreed with arguments for a light public regulatory touch. The Clinton Administration formally declared a self-regulatory approach to the internet,^[217] and subsequent administrations tended to follow suit.^[218] But as with hydrogen, government regulation played an important complementary role, and, many argue, should be even more active within the internet regulatory sphere today.^[219]

1. A Brief History of Internet Self-Regulation.

To many experts, as the seemingly impossible alien being called the “Internet” began to emerge from its military roots, there was no question that this strange and endlessly sprawling beast of an industry required self-regulation. For example, David Johnson and David Post argued that “[g]lobal computer-based communications cut across territorial borders, creating a new realm of human activity and undermining the feasibility—and legitimacy—of laws based on geographic boundaries.”^[220] Indeed, the internet emerged primarily with the support of self-regulation, but courts and legislation also played an important role throughout the development of this critical and ultimately sprawling good.

The standards that allowed the internet to flourish fall decidedly within the realm of self-regulation, although they originated within a quasi-governmental body. Specifically, in 1968, the university-based computer science contractors comprising the Advanced Research Projects Agency (ARPA, called the Defense Advanced Research Projects Agency, or DARPA, as of 1972) began a “nomadic, collegial, open and consensus-based process” for developing standards that would define an information network that would eventually come to be known as the internet.^[221] Most members of this “Network Working Group” were graduate students. This group eventually morphed into the IAB—still primarily composed of “self-selected” private individuals—and in 1986 designated an organization called the Internet Engineering Task Force (IETF) to do much of its work; the IAB reserved for itself a review and appellate role.^[222] Specifically, the IAB tasked the IETF with “the general responsibility for making the Internet work and for the resolution of all short- and mid-range protocol and architectural issues required to make the internet function effectively.”^[223] In 1992, the IETF opened up its standards process in response to criticism of relatively closed decision-making by its tight group of engineer members, and by this time it was recognized as the “primary venue for internet standards development.”^[224]

The IETF continues to serve as the primary internet standards-setting organization, although it is not formally characterized as a “voluntary consensus standards body” by the Office of Management and Budget and the Standards Development Organization Advancement Act of 2004.^[225] The IETF fails to meet this definition because it does not have formal stakeholder or interest group balance requirements, instead relying on a fully open process in which “all interested parties have an opportunity to participate.”^[226]

2. The Internet’s Foreshadowing of the Benefits and Challenges of Self-Regulation in Emerging Industries.

As would be shown for hydrogen more than twenty years later, the growth of the internet and associated self-regulatory standards demonstrated the importance—and limitations—of self-regulation in several distinct areas, including the benefits of regulatory competition among industry actors and SSOs, the coordination of standards by SSOs, and substantive and jurisdictional certainty and flexibility.

a. Competing Standards. As noted above in the hydrogen context, self-regulation can benefit emerging industries in its allowance of competition among numerous competing standards developed by industry actors and associations hoping to enable a successful new industry. Under the right circumstances—adequate coordination and information sharing, among others—the “best” standards (e.g., those that allow for growth of the new industry while also adequately addressing public concerns) may rise to the top.

The private standard that forms one of the key layers of the internet is the TCP/IP created by the quasi-SSO IETF’s predecessor IAB. TCP/IP allows individuals to communicate through "standardized packets’’ rather than direct communication that would require specified routes.^[227] This standard was itself the subject of early competition between standards-setting bodies. The ISO had created its own, competing protocol for internet communication when IAB introduced TCP/IP. Indeed, the IAB had considered attempting to join the ISO rather than maintaining its own standard-setting process through the IETF, but the ISO’s rejection of TCP/IP (the standard that ultimately won out) in part caused the IAB to remain independent of other standard-setting bodies.^[228] The IAB’s maintenance of independence, and its subsequent development of a standard-setting process that allowed every interested individual to participate, perhaps enabled more healthy competition and discussion within the internet standards-setting process than would have occurred through a more formal “balanced” or corporatist process, in which different interest groups must have specific representation.

The other unique aspect of internet standards is that due to the internet’s defiance of traditional jurisdictional control—i.e., its crossing of boundaries and accessibility to all—literally any group of individuals can create new standards, albeit informal ones. Indeed, in some cases the IETF finds itself competing with the broader user community in its standards development. For example, while the IETF created a “a minimum set of guidelines for Network Etiquette (Netiquette),” individuals and small groups created their own “Netiquette” through FAQs, which are documents on the web that “attempt to distill some Internet wisdom, or Internet norms, for newcomers.”^[229] This type of highly competitive, individual volunteer-driven standard setting has likely contributed to the success of the internet while also exacerbating its challenges, including, for example, the “vigilante” justice to which internet users sometimes resort.^[230] Indeed, more recently, some of the major competing web users’ approach to false news and harmful posts has elicited a great deal of international attention, with national efforts to address these problems often faltering.^[231] Although successful in only limited ways, shared self-regulatory approaches have emerged to somewhat curb the massive expansion of harmful information on the web.

b. The Critical Nature of Coordinated Standards. Even more so than hydrogen, which required simultaneous development of standards from the production facility to pipelines, distribution lines, and end users, the internet represented a critical need for an exceedingly complex array of standards, any of which could fail in the absence of sister standards for another component of the network. As Philip Weiser describes the internet, it consists of four layers, including content (websites), the application layer that allow internet communications applications (browser software and media players), the logical layer allowing communication among users, and the physical layer of wires and other technology that carry stored information to the user and allow users to access the internet.^[232] All of these layers are themselves quite complex. To name just a few of the complexities, the internet required the development of standards or guiding norms in all the following areas:

[W]eb site terms of use; behavioral norms of virtual chat rooms and discussion groups; network administration guidelines; listserv moderator filtering; Internet service provider contracts; Usenet voting procedures; local area network acceptable use policies; newsgroup frequently-asked question files; decisions of virtual magistrates; help manners and programmers’ manuals for multiuser dimensions; the code embedded in browsers, servers, and digital content; and the technical protocols that enable intra- and internetwork communication.^[233]

Even if several government agencies could have worked together to simultaneously write these standards, or one umbrella agency could have identified all these standards-based needs and written them at one time in whole cloth, it was unlikely that such efforts would have fully enabled an effective grid. Indeed, it was the highly technical nature of the grid, its defiance of any traditional political boundaries, and its limitless global reach that led courts, regulatory agencies such as the FCC, and many theorists to argue that self-regulation—or some other form of non-traditional governmental control—would be a key means of governing the internet. Yet even self-regulation could have been inadequate in terms of supporting internet innovation, particularly if different, competing standards emerged that prohibited the millions of networks within networks that make up the internet from connecting to each other or allowing the free flow of information among networks.

For this unwieldy, highly technical agglomeration of hardware, software, codes, wires, and the like—all in need of simultaneous development of standards—the IETF played a key role, providing a centralized, coordinating body from which the bulk of internet standards emerged.^[234]

c. Substantive and Jurisdictional Certainty and Flexibility. Beyond coordination challenges, the internet represented a classic case of a need for both regulatory certainty—particularly given the deep path dependence of internet standards—and flexibility as the internet rapidly evolved. Path dependence arises when “past decisions regarding product or standard choice will dictate the choices made in the future,” and self-regulation theorists identify path dependence as one of the features that tends to drive self-regulation.^[235] The internet had strong path-dependence features, as with, for example, the choice of languages to use for web programming.^[236] The cementing of a substantive public or private standard that would be broadly accepted and would endure as an accepted practice in the industry was therefore critical.

The need for jurisdictional certainty—in this case a new form of jurisdiction—was also central to the development of the internet. Hydrogen remains jurisdictionally uncertain, largely because a true hydrogen economy will require interstate pipelines for which Congress has not yet specified jurisdictional control over siting or construction.^[237] But the internet represented even more uncertainty with respect to who should or even could regulate the internet in the realm of public governance. Information on the internet can cross hundreds of international boundaries within nanoseconds, leading numerous early cybertheorists and courts to conclude that national control over the internet was doomed to fail.^[238] In 1997, the Southern District of New York observed that “[t]he Internet is a network of networks—a decentralized, self-maintaining series of redundant links among computers and computer networks” and concluded that “[n]o organization or entity controls the Internet; in fact, the chaotic, random structure of the Internet precludes any exercise of such control.”^[239] The court nonetheless determined that laws, such as the Commerce Clause, could reasonably apply to the internet and found that a New York law prohibiting the use of a computer to distribute obscene material impermissibly violated this clause by regulating conduct outside of New York’s borders.^[240]

Given the lack of any clear governmental control over the internet, private standards provided critical jurisdictional certainty in this space, with the IETF essentially initially serving as the international rule-setting body for the internet.

In summary, the internet provided early lessons that would foreshadow the importance of self-regulatory standards in numerous emerging, complex industries. At the same time, the development of standards for the internet demonstrates the challenges of self-regulation—challenges that are familiar to anyone working to develop the hydrogen economy.

C. Artificial Intelligence at the Cusp of Standards Development and Coordination

The field of AI differs in some respects from the two case studies explored thus far, although it, too, is emergent and “critical” in the sense of offering potentially world-changing benefits yet also substantial harms. AI as defined by the National Institute of Standards and Technology (NIST) as “an engineered or machine-based system that can, for a given set of objectives, generate outputs such as predictions, recommendations, or decisions influencing real or virtual environments.”^[241] Large language models such as ChatGPT, in turn, are a common and rapidly changing form of AI and involve “a large-scale probability machine that predicts what word should follow next based on the data it has been trained on.”^[242]

AI has been used for important applications in health, such as including “[p]redictors of health deterioration” within platforms that store health records, making health risk predictions at individual medical facilities, and identifying patients that should be transferred to intensive care.^[243] AI has also helped to detect fraud or money laundering in financial systems, among other important applications (in addition to trivial ones, such as advanced recommendations for website perusal and shopping destinations).^[244] The technology is so game changing that one frequently hears forecasts that it will end work as we know it.^[245]

We briefly explore AI as a case study, despite some differences from hydrogen and the internet, because it is one of the most prominent examples of an emerging industry, and its regulatory journey—both private and public—is only just beginning. In terms of its differences from our other two examples, AI as an industry is not particularly networked. It does not involve a group of disparate technologies and practices in different business sectors that require coordination in order for the industry to be effective.^[246] Many AI applications draw and connect massive amounts of data from a variety of sources, however, and AI’s use in third-party applications through plugins requires some degree of coordination.^[247] Also in contrast with an industry such as hydrogen, the emergence of AI seems inevitable, rather than depending to a meaningful degree on the existence of standards to, for example, address coordination difficulties or allay public fears.^[248]

Despite these differences from other emerging industries, as with our other case studies and additional examples explored here, AI’s trajectory will in large part depend on the effectiveness of some combination of public and private standards. For example, some experts have expressed concerns about emerging regulations that create “significant precedent,” in a negative sense, for the industry, such as deletion of entire machine learning models and algorithms by regulators.^[249] And as discussed below, there have already been bans and threatened temporary suspensions of AI products such as ChatGPT.

The role of private standards and the emerging application of existing public laws to AI shows just how important a suite of public–private rules can be to address the externalities of an emerging industry and to fill jurisdictional gaps as the industry rapidly expands. It also demonstrates the importance of both government actors and NGO actors in helping to coordinate private and public standards as they emerge, similar to the hydrogen story.

1. Regulatory Gaps and Uncertainty.

Few existing U.S. statutes or regulations directly apply to AI, and private standards are only slowly emerging. Some would argue that the relative dearth of public governance is beneficial given the highly complex and technical nature of AI that would lead to “clumsy” and quickly outdated public regulation.^[250] But in the same breath, those acknowledging the benefits of self-regulation for AI also worry that self-regulation in this area lacks adequate enforcement mechanisms and express concern that self-regulatory mechanisms would not be mandatory.^[251] Given its capacity to use and disseminate others’ information, predict crime rates in specific geographic areas and individuals’ likelihood of committing crimes, produce disinformation, and predict disease, the risks of AI with respect to privacy, misappropriation of data, bias, and perpetuation of manipulative and false information are extraordinary.^[252]

The primary means of public regulation of AI has been through the FTC, which monitors for and enforces issues such as deceptive advertising in AI (unsupported claims that it can perform in a superior way to other products, for example).^[253] The FTC has also used its jurisdiction over deceptive and misleading claims to enforce AI’s misleading of web users through practices such as “deepfakes” and “chatbots.”^[254] Further, it is enforcing companies’ use of fraudulently or misleadingly obtained data in algorithms, in some cases requiring algorithmic deletion—“the ordered deletion of computer data models or algorithms that were developed with improperly obtained data.”^[255]

Beyond the application of long-running standards for fraudulent and deceptive trade practices to this new technology, few public regulations apply to AI. This allows for potentially large negative externalities as AI grows. It also produces regulatory uncertainty, including bans and threats of bans and moratoria. For example, Italy temporarily effectively banned ChatGPT, alleging violations of European Union data protection rules.^[256] And prominent industry leaders in the United States have called on the FTC to temporarily suspend ChatGPT releases, alleging unfair and deceptive data practices.^[257]

Private standards and best practices are emerging in several different forms, although they are so new that it is not yet clear how they will be enforced or how effective they will be. The Frontier Model Forum—a group of leading AI industries—launched in July 2023 and has not yet issued standards.^[258] The private standards already published by the Forum’s major members, however, may foreshadow the likely content of consensus standards. OpenAI—the producer of ChatGPT—has developed policies for usage; content guidelines; and policies for plugins that connect OpenAI to third-party applications, among other standards.^[259] Its list of prohibited uses is the lengthiest set of standards. To name just a few examples, OpenAI disallows “the use of our models, tools, and services for illegal activity”; sexual abuse; “hateful, harassing, or violent content”; “[g]eneration of malware”; activities with “high risk of physical harm” such as “[w]eapons development”; and activities with “high risk of economic harm” such as “[g]ambling” and “[p]ayday lending.”^[260] OpenAI indicates that it will enforce these standards by asking the user of an OpenAI product to “make necessary changes” if a violation is discovered and potentially “suspending or terminating” a user’s account in the event of “[r]epeated or serious violations.”^[261]

SSOs are also beginning to take a gap-filling role, with the NIST releasing an AI Risk Management Framework in January 2023.^[262] This is a voluntary set of standards designed to “to improve the ability to incorporate trustworthiness considerations into the design, development, use, and evaluation of AI products, services, and systems” and, accordingly, “to better manage risks to individuals, organizations, and society.”^[263] This consensus-based document, informed by several workshops with public and private stakeholders, a request for information, and public comments, identifies seven central “[c]haracteristics of trustworthy AI systems,” such as systems that are “safe,” “secure and resilient,” and “privacy-enhanced.”^[264] It then defines these characteristics in detail and provides examples—at a broad level—of how the characteristics can be achieved. For instance, under the characteristic of an “accountable and transparent” AI system, the Framework notes that “[t]ransparency should consider human-AI interaction: for example, how a human operator or user is notified when a potential or actual adverse outcome caused by an AI system is detected.”^[265] The document also notes that AI actors should adjust their actions in a manner that is proportionate to potential harm, such as “when life and liberty are at stake” as a result of an AI system.^[266]

Public law directed specifically at AI is sparse, although the federal government has defined AI by statute (in the John S. McCain National Defense Authorization Act of 2019), and President Trump issued policy guidance on AI in Executive Order 13859 in February 2019, and an associated Office of Management and Budget memorandum.^[267] The memorandum emphasized a hands-off approach to AI regulation, stating:

Federal agencies must avoid regulatory or non-regulatory actions that needlessly hamper AI innovation and growth. Where permitted by law, when deciding whether and how to regulate in an area that may affect AI applications, agencies should assess the effect of the potential regulation on AI innovation and growth. While narrowly tailored and evidence-based regulations that address specific and identifiable risks could provide an enabling environment for U.S. companies to maintain global competitiveness, agencies must avoid a precautionary approach that holds AI systems to an impossibly high standard such that society cannot enjoy their benefits . . . .^[268]

The memorandum listed ten high-level principles that federal agencies should follow when considering AI regulatory or non-regulatory actions, including, for example, cost-benefit approaches, public participation in AI rulemaking, performance-based flexible regulation, and fairness and non-discrimination.^[269] A statute and executive order also guide the federal government’s use of AI.^[270]

Some older statutes directed at specific professions, such as pathology, also provide limits on the acceptable use of AI and requirements for the validation of algorithms.^[271] Beyond this, most federal governance of AI is designed to incentivize and further the growth of AI, through the creation of the National Artificial Intelligence Initiative (and a federal office by the same name) in 2020, and a $100 million National Science Foundation Investment in Artificial Intelligence Institutes.^[272]

2. Coordination of Emerging Standards.

The standards emerging for AI are even more nascent than for hydrogen, and the coordination of standards seen in the hydrogen context has yet to occur. For hydrogen, industry associations have begun to coalesce around consensus standards, and the DOE plays a coordinating role, highlighting gaps in the standards.^[273] The AI industry appears to be at an earlier point, with individual AI actors setting their own rules and only just beginning to coordinate themselves.^[274] Further, although the Office of Management and Budget has issued a guidance document for how public regulation of AI should proceed, there does not appear to be a coordinating champion for public and private AI standards.^[275] The differences in approaches to private standards in the AI context—and differing opinions about the extent of public governance needed—were highlighted in a May 2023 congressional hearing on ChatGPT.^[276]

Sam Altman, the Chief Executive Officer of OpenAI, testified extensively about the internal protocols that the company follows to protect privacy, police disinformation, protect children’s safety, ensure cybersecurity, improve accuracy, and deploy ChatGPT products safely. Altman cited to OpenAI’s Usage Policies and noted that OpenAI uses “a combination of automated detection systems and human review to detect potentially violating behavior in order to warn users or take enforcement actions.”^[277] He also, however, emphasized that “regulation of AI is essential” and noted that OpenAI is working to apply the NIST AI principles to its models.^[278] Christina Montgomery, Chief Privacy and Trust Officer for IBM, focused more on the company’s self-regulation, emphasizing its creation of an AI Ethics Board, for example, while Sam Altman listed the specific steps taken to minimize risk posed by OpenAI.^[279]

While operating with little coordination of standards on the domestic front, the U.S. AI community is also endeavoring to harmonize its emerging standards with international ones—an effort reminiscent of the early days of the internet. Brookings, which has led some of the efforts in this space, notes its own forum for AI cooperation formed in collaboration with the Centre for European Policy Studies, as well as G-7 Leaders Communique on AI and work on AI governance through “the US–EU Trade and Technology Council (TTC), the Global Partnership in AI (GPAI), [and] the Organization for Economic Co-operation and Development (OECD).”^[280]

Overall, AI—one of the most nascent yet potentially transformative industries at the time of this writing—highlights well the role of standards in innovation. The U.S. government is pulling many of the traditional levers that enable and support innovation, such as forming national R&D centers through the National Science Foundation and creating a federal office to issue incentives and support for AI. At the same time, the government is engaged in the delicate dance of considering its own new standards to mitigate the risks of AI—thus supporting public confidence in this emerging area—while recognizing that too many standards could backfire and hamper innovation. In the meantime, self-regulation is emerging, in part, due to recognition of the chilling effect that highly publicized AI risks could have on the industry and the real threat of bans, as seen in Italy. Efforts to coordinate these standards through, for example, the Frontier Model Forum and government- and industry-led coordination efforts are still relatively young. The degree to which a suite of complementary public and private standards will support productive, transformative use of AI and curb potentially large negative harms—and effectively harmonize international standards to support the growth of the industry—remains to be seen.

V. Lessons Learned

The observation that private standards development has tended to follow commercialization of technologies—made by the president of the technical branch of the leading U.S. standards organization, ASME Standards Technology LLC—is an intriguing and frankly surprising one.^[281] How, after all, could an industry develop so fully—to the point of full commercialization—without standards guiding it and reassuring the public of the industry’s efficacy and safety? Hydrogen is perhaps one of the strongest exceptions to this rule, demonstrating the critical role of standards in an industry that has barely reached the pilot stage in many contexts. Failed efforts by California to produce a viable “Hydrogen Highway” are the result of many factors, but a lack of clear public or private standards for hydrogen fueling station safety may have proven a leading impediment, spurring the private NFPA and federal agencies to develop clear permitting guides and videos for hydrogen fueling stations in the more recent effort to scale up hydrogen.

With hydrogen, standards appear to be one determinant of whether the industry will flourish or, once again, fall by the wayside as a minor player in the energy sector. In the case of the internet, standards, too, were critical for issues such as interoperability. For AI, the industry appears to have few impediments to growth; standards are essential not for its emergence, but to temper potentially harmful impacts and to harness this technology to produce meaningful benefits—i.e., medical breakthroughs rather than merely law essays re-formulated in the style of great jurists.

When is self-regulation critical to the emergence of a nascent industry, and what types of standards prove most effective in pushing such an industry toward a commercial, scaled-up, and net-positive state? A great deal of additional empirical analysis is essential to fully answer this question, but we offer initial thoughts here.

Prior analyses of self-regulation suggest that industries with strong network externalities and path dependence, among other market factors, will gravitate toward self-regulation. With respect to network externalities, in which “the value of a product to consumers is heavily dependent upon its acceptance by other consumers,” the internet clearly reflects this feature.^[282] The more individuals connect to the internet, the more information is accessible and shared; the more localized networks that connect with others (perhaps most efficiently through shared, commonly understood standards), the more information flow there will be. With hydrogen, too, more consumer demand for hydrogen vehicles and associated fueling stations benefits individual hydrogen vehicle users. Path dependence, in which a choice of web programming language or type of hydrogen fueling station, can also affect the future feasibility of an industry, which tends to call for self-regulation.

Our exploration of the emerging hydrogen industry and associated standards, and the now mature internet, suggests several other important lessons about industry features—particularly in emerging industries—that tend to suggest a need for self-regulation complemented by public governance. First, and perhaps foremost, is the importance of standards in a technical, networked industry that the public perceives to be highly risky or highly novel. In the case of hydrogen, industry actors in the United Kingdom and United States, among many other jurisdictions, have launched efforts to technically address risk in order to support standards development and assuage public fears, with U.K. gas distribution companies funding collective risk laboratories and the ASME producing numerous technical papers to inform hydrogen pipeline standards.^[283] In other, simpler industries, with fewer supply chain components and end uses and more familiar technologies, perhaps the industry, the public at large, and regulators are more willing to watch the industry emerge, learning of risks along the way and developing standards ex ante. Not so, it seems, in the hydrogen space.

Another feature of emerging industries that seemingly calls for relatively robust self-regulation is a more complex version of the network externalities story. Beyond involving a product that becomes more valuable the more widespread and adopted it becomes, emerging industries that seem to have benefited from—or even depended upon—industry standards involve a complex physical network with mutually dependent components. For hydrogen, these are the production facilities, pipelines, storage facilities, distribution lines, fueling stations, and vehicles. For the internet, networked infrastructure is cyberspace—an “interdependent network of information systems infrastructures including . . . telecommunications networks, computer systems, and embedded processors and controllers.”^[284]

Beyond the need for standards during the commercialization process, hydrogen and the internet inform questions regarding the types of standards that can be most effective in supporting this process. We have already noted the importance of standards buttressed by powerful technical proof. Beyond this need is the critical nature of standards coordination in an industry characterized by a complex network of production, transportation, distribution, and end-use infrastructure and operation. For hydrogen, the Fuel Cell & Hydrogen Energy Association is pushing to identify and fill areas currently lacking in standards guidance and to mesh standards development processes toward a more coordinated end goal. Government actors, in turn, such as the National Renewable Energy Laboratory and Sandia National Laboratories, are helping to inform industry actors and the public of the relevant public and private standards—thus easing the permitting of new hydrogen infrastructure—and to support standards development. Additionally, individual standards development organizations are prodding others to improve hydrogen standards and fill gaps, as demonstrated by the CGA asking the NFPA to clarify some of its hydrogen storage standards and creating an advisory document while awaiting this clarification.

As is occurring with the now-mature internet, as emerging industries such as AI grow and change over time, governance will evolve to represent a more complex mix of private and public regulations—often with public regulations incorporating many private standards by reference but also including new publicly formed directives. And the debate between the merits of private and public standards will become more intense, as seen by the growing cacophony of voices concerned about self-regulation of internet content and speech and AI’s appropriation of others’ work, among many other issues. Ultimately, the standards that seem most likely to expand within the public sphere are those defined by Larry Lessig as “regulating” standards—those that “limit liberty within” an activity to “advance a regulatory end,” such as public safety.^[285] Indeed, Lessig noted that regulatory standards tend to be top-down.^[286]

Despite the likelihood of more top-down public regulation of emerging industries as they grow—with the internet as a prime example—in some cases there are likely to be strong bottom-up additions of regulating standards emanating pioneered by SSOs and SROs, particularly if industry members of these organizations become concerned about unduly strict public regulation or interconnected risks that could cause the whole system to fail. In the case of hydrogen, particularly given stubborn public perceptions of risk in the form of fire and explosion, one or several high-profile incidents could doom the industry, as could the failure of a critical pipeline connected multiple producers to end-users. This could lead industry to continue to play an active role both in curating the coordinating standards (those that allowed the industry to exist in the first place) and the regulating standards, which improve safety, public perception, and other aspects of the now-functioning industry.

VI. Conclusion

An interesting yet high-stakes experiment in innovation is currently playing out within the United States and globally. It is unclear at this juncture whether extensive global efforts to create a viable hydrogen economy will follow the path of California’s doomed Hydrogen Highway or meet a happier fate. And we cannot now know just how transformative AI will be, and the extent and degree of the risks of this massive technological change, let alone how well public and private actors will respond to these risks with regulatory governance. Will a global clean hydrogen economy emerge and become part of the solution to climate change? Will AI become a central, productive, and effectively regulated part of most aspects of human life? Much of this depends on the success of the technological development currently supported by a growing suite of industry standards.

The role of standards in the future story of these industries is a critical one and provides lessons well beyond the limited spheres explored here, be it the development of advanced medical treatments and vaccinations, self-driving vehicles, and other potentially beneficial technologies currently only on the cusp of commercialization. Ultimately, economics, consumer confidence, and dynamics in other industries are likely to determine the fate of these types of emerging industries. But private standards supported and complemented by public governance will have played a critical role along the winding path of this nascent industry, just as they have and will continue to do for numerous nascent industries.

---

1. Oscar Gelderblom & Regina Grafe, The Rise and Fall of the Merchant Guilds: Re-thinking the Comparative Study of Commercial Institutions in Premodern Europe, 40 J. Interdisc. Hist., 477, 483 (2010).

2. See infra Part II.

3. For examples of the relatively few sources that have focused specifically on self-regulation and topics relating to emerging industries, see infra note 5.

4. Industrial policy is the full suite of factors that enable industry to thrive, such as support for research, access to foreign markets, beneficial taxation measures, and the like. Michael Maibach, The Industrial Policy Cycle, Stan. L. & Pol’y Rev., Fall 1993, at 23, 23–26, 28. The literature typically views industrial policy as emanating from Congress and the President—think, for example, of the CHIPS Act of 2022 § 102, 15 U.S.C. § 4651 (supporting the semiconductor industry). But we view industrial policy as decidedly within both camps—public policy and self-regulation—and explore the ideal balance of the two.

5. There are a few exceptions to this siloing in the literature, though most of them focus on the relationship between self-regulation and innovation only within one particular sector. See, e.g., Cristie Ford, Innovation and the State: Finance, Regulation, and Justice 44 (2017); Timothy Craig Allen, Regulating Artificial Intelligence for a Successful Pathology Future, 143 Archives Pathology & Lab’y Med. 1175, 1176 (2019); Molly Cohen & Arun Sundararajan, Self-Regulation and Innovation in the Peer-to-Peer Sharing Economy, 82 U. Chi. L. Rev. Online 116, 118, 121–22, 129 (2015), https://chicagounbound.uchicago.edu/cgi/viewcontent.cgi?article=1039&context=uclrev_online [https://perma.cc/4N56-7URJ]; Christodoulos Stefanadis, Self-Regulation, Innovation, and the Financial Industry, 23 J. Regul. Econ. 5, 18 (2003); Paul Moritz Wiegmann et al., Competing Standard-Setting Organizations: A Choice Experiment, Rsch. Pol’y, 2022, at 1, 11–12.

6. . See, e.g., Mancur Olson, The Logic of Collective Action: Public Goods and the Theory of Groups 53–57, 132–33, 138, 142–43 (1965) (exploring the conditions under which groups most effectively organize); Andrew A. King & Michael J. Lenox, Industry Self-Regulation Without Sanctions: The Chemical Industry’s Responsible Care Program, 43 Acad. Mgmt. J. 698, 702 (2000).

7. . Ian Ayres & John Braithwaite, Responsive Regulation: Transcending the Deregulation Debate (1992); Cary Coglianese & Evan Mendelson, Meta-Regulation and Self-Regulation, in The Oxford Handbook of Regulation 146 (Robert Baldwin et al. eds., 2010); Cary Coglianese & Jennifer Nash, Motivating Without Mandates? The Role of Voluntary Programs in Environmental Governance, in 2 Elgar Encyclopedia of Env’t L., Decision Making in Environmental Law 237 (LeRoy C. Paddock et al. eds., 2016); Peter Grajzl & Peter Murrell, Allocating Lawmaking Powers: Self-Regulation vs. Government Regulation, 35 J. Compar. Econ. 520 (2007).

8. See, e.g., Amanda C. Leiter, Fracking, Federalism, and Private Governance, 39 Harv. Env’t L. Rev. 107 (2015); Aseem Prakash & Matthew Potoski, Voluntary Environmental Programs: A Comparative Perspective, 31 J. Pol’y Analysis & Mgmt. 123 (2012); Michael P. Vandenbergh & Jonathan M. Gilligan, Beyond Politics: The Private Governance Response to Climate Change (2017).

9. Amy L. Stein, Reconsidering Regulatory Uncertainty: Making a Case for Energy Storage, 41 Fla. St. U. L. Rev. 697 (2014); Sarah E. Light, Precautionary Federalism and the Sharing Economy, 66 Emory L.J. 333 (2017).

10. See, e.g., Edwin Mansfield, Patents and Innovation: An Empirical Study, 32 J. Inst. Mgmt. Scis. 173, 179–80 (1986); Brett Frischmann, Innovation and Institutions: Rethinking the Economics of U.S. Science and Technology Policy, 24 Vt. L. Rev. 347, 356, 366–67, 398 (2000); Daniel J. Hemel & Lisa Larrimore Ouellette, Beyond the Patents—Prizes Debate, 92 Tex. L. Rev. 303, 316 & n.40 (2013); Zhongmin Wang & Alan Krupnick, A Retrospective Review of Shale Gas Development in the United States: What Led to the Boom? 34 (Res. for the Future, Working Paper, RFF DP 13-12, 2013), https://media.rff.org/documents/RFF-DP-13-12.pdf [https://perma.cc/BZ2D-AJ7E]; John M. Golden & Hannah J. Wiseman, The Fracking Revolution: Shale Gas as a Case Study in Innovation Policy, 64 Emory L.J., 955, 995–98 (2015); Yochai Benkler, Peer Production, the Commons, and the Future of the Firm, 15 Strategic Org. 264, 264–66 (2017).

11. Coglianese & Mendelson, supra note 7, at 162–63; Ayres & Braithwaite, supra note 7, at 101–06.

12. See Morgan Ricks et al., Networks, Platforms, and Utilities: Law and Policy (2022).

13. Coglianese & Mendelson, supra note 7, at 146, 151, 157.

14. Id. at 150, 152; John W. Maxwell et al., Self-Regulation and Social Welfare: The Political Economy of Corporate Environmentalism, 43 J.L. & Econ. 583, 587 (2000). When government regulators maintain some form of oversight and control of firm- or individual-level regulation, regulatory scholars call that “meta-regulation.” Coglianese & Mendelson, supra note 7, at 147–48. Such meta-regulation can occur in all different shades of gray: sometimes government regulators incorporate private industry standards by reference; other times government regulators maintain some kind of loose process of review over self-regulatory organizations. Emily S. Bremer, Teaching Guide, Technical Standards Meet Administrative Law: A Teaching Guide on Incorporation by Reference, 71 Admin. L. Rev. 315, 322, 324–25 (2019) (discussing processes of incorporation by reference, wherein government regulators adopt privately-promulgated standards); Emily Hammond, Double Deference in Administrative Law, 116 Colum. L. Rev. 1705, 1731 (2016) (discussing problems that emerge in government oversight of self-regulatory organizations’ activities).

15. About ANSI, Am. Nat’l Standards Inst., https://www.ansi.org/about/introduction [https://perma.cc/KTA8-FMFZ] (last visited Sept. 28, 2024); Detailed Overview, Am. Soc’y for Testing & Materials Int’l, https://www.astm.org/about/overview/detailed-overview.html [https://perma.cc/Z5FW-SNAF] (last visited Sept. 29, 2024); About ASME, Am. Soc’y of Mech. Eng’rs, https://www.asme.org/about-asme [https://perma.cc/R39G-5DSD] (last visited Sept. 29, 2024).

16. JoAnne Yates & Craig N. Murphy, Engineering Rules: Global Standard Setting Since 1880 at 127, 130 (2019).

17. Hammond, supra note 14, at 1706–07, 1715–16; Saule T. Omarova, Wall Street as Community of Fate: Toward Financial Industry Self-Regulation, 159 U. Pa. L. Rev. 411, 417 (2011) [hereinafter Omarova, Community of Fate].

18. Kenneth W. Abbott et al., Theorizing Regulatory Intermediaries: The RIT Model, 670 Annals Am. Acad. Pol. & Soc. Sci. 14, 14–15 (2017); see also Daniel E. Walters, Reclaiming Regulatory Intermediation for the Public, 86 Law & Contemp. Probs. 157, 157, 162 (2023).

19. See, e.g., Cultivate Credibility: Standards and Solutions for the Cannabis and Hemp Industry, ASTM Int’l Cannabis Standards & Servs., https://www.astmcannabis.org [https://perma.cc/283H-GUME] (last visited Sept. 9, 2024); About Us, Nat’l Cannabis Indus. Ass’n, https://thecannabisindustry.org/about-us/ [https://perma.cc/Q67T-FJKS] (last visited Oct. 11, 2024); Roncevert Ganan Almond, After the Max: Rebuilding U.S. Aviation Leadership, Va. J. Int’l L. Online, 2019, at 1, 6–8 (describing and critiquing the rise of self-regulation in the aviation industry); Robert Heidt, Industry Self-Regulation and the Useless Concept “Group Boycott,” 39 Vand. L. Rev. 1507, 1508–09 (1986) (summarizing the self-regulation cases ranging from regulation of professionals such as doctors and chiropractors to consumer products and sports players and gear); Sarah B. Schindler, Following Industry’s LEED: Municipal Adoption of Private Green Building Standards, 62 Fla. L. Rev. 285, 303–04 (2010); Joanna R. Schacter, Note, Delegating Safety: Boeing and the Problem of Self-Regulation, 30 Cornell J.L. & Pub. Pol’y 637 (2021); The Critical Connection Between Consumers & Standards, ANSI, https://www.ansi.org/outreach/consumers/consumers-standards [https://perma.cc/H387-NEFV] (last visited Dec. 15, 2024); Genevieve Diesing, The Essential Benefits of ISO Standards for Manufacturers, Quality Mag. (Aug. 25, 2023), https://www.qualitymag.com/articles/97509-the-essential-benefits-of-iso-standards-for-manufacturersmission-values/ [https://perma.cc/77S3-N4AN] (last visited Oct. 11, 2024).

20. See Douglas C. Michael, Federal Agency Use of Audited Self-Regulation as a Regulatory Technique, 47 Admin. L. Rev. 171, 174 n.8, 175 n.12, 181 (1995) (noting that in highly technical contexts the self-regulator possesses “superior knowledge of the subject compared to the government agency”); Omarova, Community of Fate, supra note 17, at 434 (emphasizing self-regulation’s advantages relative to government regulation where technical complexities make it difficult for government to keep up).

21. See David P. McCaffrey & David W. Hart, Wall Street Polices Itself: How Securities Firms Manage the Legal Hazards of Competitive Pressures 66–91 (1998); Omarova, Community of Fate, supra note 17, at 417; Saule T. Omarova, Rethinking the Future of Self-Regulation in the Financial Industry, 35 Brook. J. Int’l L. 665, 676–77, 693 (2010) [hereinafter Omarova, Rethinking the Future].

22. Benjamin P. Edwards, The Dark Side of Self-Regulation, 85 U. Cin. L. Rev. 573, 579, 581–83, 586 (2017); Omarova, Community of Fate, supra note 17, at 417. Some argue that this oversight is no longer so loose and that the SEC is becoming an embedded “meta-regulatory” of NYSE and FINRA. See, e.g., id. at 464–66, 485 (describing the SEC’s oversight of NYSE and FINRA as “comprehensive oversight”).

23. What Is Financial Technology (FinTech)? A Beginner’s Guide, Colum. Eng’g: Boot Camp, https://bootcamp.cvn.columbia.edu/blog/what-is-fintech/ [https://perma.cc/QAS6-RL6M] (last visited Sept. 9, 2024).

24. See, e.g., Chris Brummer & Yesha Yadav, Fintech and the Innovation Trilemma, 107 Geo. L.J. 235, 304–06 (2019); Matthias Lehmann, Global Rules for a Global Market Place? – Regulation and Supervision of Fintech Providers, 38 B.U. Int’l L.J. 118, 132 (2020).

25. See Timothy G. Massad & Howell E. Jackson, How to Improve Regulation of Crypto Today—Without Congressional Action—and Make the Industry Pay for It 7, 13–14 (Hutchins Ctr., Working Paper No. 79, Oct. 2022), https://www.brookings.edu/wp-content/uploads/2022/10/WP79-Massad-Jackson-updated-2.pdf [https://perma.cc/3LZJ-HSNM]; Eric Alston, Digital Currency Industry Self-Regulation: Not All Consensus Is Automatic, Va. J.L. & Tech., Summer 2023, at 1, 5, 8, 13–15, 36, 39, 42.

26. Coglianese & Nash, supra note 7, at 238, 248–49.

27. Coglianese & Mendelson, supra note 7, at 154; King & Lenox, supra note 6, at 699–700, 700 tbl.1.

28. Prakash & Potoski, supra note 8, at 128–29.

29. Vandenbergh & Gilligan, supra note 8, at 14–15, 119.

30. Coglianese & Mendelson, supra note 7, at 155–56.

31. These include the North American Electricity Reliability Corporation (NERC) and the North American Energy Standards Board (NAESB). Hammond, supra note 14, at 1741; Alexandra Klass, Joshua Macey, Shelley Welton & Hannah Wiseman, Grid Reliability Through Clean Energy, 74 Stan. L. Rev. 969, 993 (2022); Joshua C. Macey, Shelley Welton & Hannah Wiseman, Grid Reliability in the Electric Era, 41 Yale J. on Regul. 164, 203 (2024). A meta-regulator, the Federal Energy Regulatory Commission, oversees NERC’s standard-setting and enforcement, with a healthy dose of deference to NERC and private sub-entities of NERC called Regional Entities that do much of the organization’s work. Id. at 174. A meta-regulator is a government entity that regulates a self-regulatory organization. Omarova, Community of Fate, supra note 17, at 482 & nn.262–63.

32. See, e.g., Daniel E. Walters & Andrew N. Kleit, Grid Governance in the Energy-Trilemma Era: Remedying the Democracy Deficit, 74 Ala. L. Rev. 1033, 1036, 1042 (2023); Hari M. Osofsky & Hannah J. Wiseman, Hybrid Energy Governance, 2014 U. Ill. L. Rev. 1, 7 (2014); Shelley Welton, Rethinking Grid Governance for the Climate Change Era, 109 Calif. L. Rev. 209, 261–62 (2021).

33. Standards, Am. Petrol. Inst., https://www.api.org/products-and-services/standards/ [https://perma.cc/NA6F-WJQV] (last visited Sept. 29, 2024).

34. Am. Petrol. Inst., 2022 API Publications Catalog (2022), https://www.api.org/-/media/files/publications/2022-catalog/2022-pubs-catalog-for-website.pdf [https://perma.cc/GCU3-UW74]; Leiter, supra note 8, at 128; Hannah Jacobs Wiseman & Francis Gradijan, Regulation of Shale Gas Development, Including Hydraulic Fracturing (U. Tulsa Legal Stud. Rsch. Paper No. 2011-11, 2012), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1953547 [https://perma.cc/WGN7-AU67].

35. See infra Part III; Oleksiy Taterenko et al., Hydrogen State of the Union: Where We Stand in 2024, Rocky Mtn. Inst. (Apr. 18, 2024), https://rmi.org/hydrogen-state-of-the-union-where-we-stand-in-2024/ [https://perma.cc/3YAJ-V9W8].

36. Hydrogen Production, U.S. Dep’t of Energy, https://www.energy.gov/eere/fuelcells/hydrogen-production [https://perma.cc/LLP3-GUJ2] (last visited Nov. 15, 2024); Hydrogen Delivery, U.S. Dep’t of Energy, https://www.energy.gov/eere/fuelcells/hydrogen-delivery [https://perma.cc/86EW-NXLM] (last visited Nov. 15, 2024). With proper modifications and safety precautions, hydrogen can be burned or used in lieu of natural gas and other fuels in a residential stove or furnace, industrial boiler or process, or power plant. See, e.g., Leeds Climate Comm’n, Hydrogen Conversion: Potential Contribution to a Low Carbon Future for Leeds 2019, https://www.leedsclimate.org.uk/hydrogen-conversion-potential-contribution-low-carbon-future-leeds [https://perma.cc/E9NN-GW2R] (last visited Nov. 15, 2024) (describing the H21 Leeds City Gate project to deliver hydrogen rather than natural gas to a variety of end users).

37. See, e.g., Grace Heusner et al., Defining and Closing the Hydraulic Fracturing Governance Gap, 95 Denv. L. Rev. 191, 195, 212 (2017) (arguing for more local control over hydraulic fracturing for oil and gas in light of a failure to address some risks at the state and federal levels); Macey, Welton & Wiseman, supra note 31, at 234–38 (documenting problems with electricity reliability regulation). There are no federal provisions indicating whether the federal government or states will permit interstate hydrogen pipelines or import and export terminals, making it difficult to coordinate the infrastructure buildout necessary to scale the industry. Austin R. Baird et al., Sandia Nat’l Lab’ys, Federal Oversight of Hydrogen Systems 10–12 (2021), https://energy.sandia.gov/wp-content/uploads/2021/03/H2-Regulatory-Map-Report_SAND2021-2955.pdf [https://perma.cc/JC54-KFXK]. This may change in the near future, since part of Senator Joe Manchin’s deal to secure passage of the Inflation Reduction Act was a commitment to a laundry list of “permitting reforms” to prioritize energy infrastructure development. Part of this laundry list was a broad proposal to give the FERC authority to permit hydrogen pipelines, storage, and import/export facilities. See Omar Samji et al., A&O Shearman, After the Inflation Reduction Act: Permitting Reforms to Expedite EnerInfrastructure Projects, Lexology (Aug. 29, 2022), https://www.lexology.com/library/detail.aspx?g=e8a85a91-cb8e-4a7b-81d6-f889c00605fa [https://perma.cc/Z6J4-3MT6]. However, as we write, there is a growing uncertainty whether Manchin’s proposals will in fact become law. See Rachel Frazin & Zack Budryk, Are Permitting Reform Talks Under Threat?, Hill (Aug. 3, 2023, 6:18 PM), https://thehill.com/newsletters/energy-environment/4136651-are-permitting-reform-talks-under-threat/ [https://perma.cc/QND5-4ATG].

38. Safe Use of Hydrogen, U.S. Dep’t of Energy, https://www.energy.gov/eere/fuelcells/safe-use-hydrogen [https://perma.cc/856F-E9YF] (last visited Nov. 16, 2024).

39. Jenna Barron, Anthropic, Google, Microsoft, and OpenAI Form Group Dedicated to Safe Development of Frontier AI Models, SD Times (July 28, 2023), https://sdtimes.com/ai/anthropic-google-microsoft-and-openai-form-group-dedicated-to-safe-development-of-frontier-ai-models/ [https://perma.cc/QU92-EQV4].

40. Frontier Model Forum, OpenAI (July 26, 2023), https://openai.com/index/frontier-model-forum/ [https://perma.cc/G2LF-6BD2].

41. Bletchley Park, The World Wants to Regulate AI, But Does Not Quite Know How, Economist (Oct. 24, 2023), https://www.economist.com/business/2023/10/24/the-world-wants-to-regulate-ai-but-does-not-quite-know-how [https://perma.cc/LX86-ANQR]; François Candelon et al., AI Regulation Is Coming, Harv. Bus. Rev. (Sept.–Oct. 2021), https://hbr.org/2021/09/ai-regulation-is-coming [https://perma.cc/JMD3-3T4E].

42. See infra Section IV.B.

43. Ford, supra note 5, at 25.

44. Ayres & Braithwaite, supra note 7, at 4, 101, 103.

45. Jody Freeman, Collaborative Governance in the Administrative State, 45 UCLA L. Rev. 1, 22 (1997).

46. Omarova, Community of Fate, supra note 17, at 416–17.

47. See generally Richard F. Hirsh, Power Loss: The Origins of Deregulation and Restructuring in the American Electric Utility System (1999).

48. See supra Section II.A.

49. Jodi L. Short & Michael W. Toffel, Making Self-Regulation More than Merely Symbolic: The Critical Role of the Legal Environment, 55 Admin. Sci. Q. 361, 361 (2010).

50. The reputational driver is complex. Much of it, however, assumes that industry self-regulation is particularly likely to emerge when the misbehavior of any one firm within the industry could ruin the reputation of the entire firm. Prakash & Potoski, supra note 8, at 128; Omarova, Community of Fate, supra note 17, at 446.

51. King & Lenox, supra note 6, at 698 (noting concern that “a few poor performers can lead to environmental regulation of the industry as a whole”); Maxwell et al., supra note 14, at 584, 587, 608–09 (describing “strategic self-regulation that preempts government action”).

52. See supra note 20 and accompanying text; Omarova, Community of Fate, supra note 17, at 433–34 (“[L]everaging the industry actors’ relatively greater abilities to understand and analyze increasingly complex and overwhelmingly voluminous financial information offers a major potential benefit from the perspective of regulatory efficiency and efficacy.”).

53. See generally Neil Gunningham et al., Shades of Green: Business, Regulation, and Environment (2003).

54. Michael, supra note 20, at 181–82; Omarova, Community of Fate, supra note 17, at 434–35.

55. Ruthanne Huising & Susan S. Silbey, Accountability Infrastructures: Pragmatic Compliance Inside Organizations, 15 Regul. & Governance S40–S41 (2021) (describing tailoring of self-regulation to specific circumstances).

56. Ayres & Braithwaite, supra note 7, at 4, 110–11.

57. Wiegmann et al., supra note 5, at 1–2 (citing research on “the importance of standards” in innovative, networked industries); Stefanadis, supra note 5, at 5–6, 21–22.

58. Coglianese & Mendelson, supra note 7, at 160.

59. Shanti Gamper-Rabindran & Stephen R. Finger, Does Industry Self-Regulation Reduce Pollution? Responsible Care in the Chemical Industry, 43 J. Regul. Econ. 1, 3 (2013).

60. King & Lenox, supra note 6, at 700–01; Short & Toffel, supra note 49, at 364.

61. King & Lenox, supra note 6, at 701.

62. Short & Toffel, supra note 49, at 366.

63. See Michael J. Lenox & Jennifer Nash, Industry Self-Regulation and Adverse Selection: A Comparison Across Four Trade Association Programs, 12 Bus. Strategy & Env’t 343, 346–47 (2003); King & Lenox, supra note 6, at 702.

64. See Coglianese & Mendelson, supra note 7 at 160–61; see also King & Lenox, supra note 6, at 701–02. See generally Neil Gunningham, Environment, Self-Regulation, and the Chemical Industry: Assessing Responsible Care, 17 Law & Pol’y 57 (1995).

65. Coglianese & Mendelson, supra note 7, at 161.

66. See Olson, supra note 6, at 52–54.

67. Omarova, Rethinking the Future, supra note 21, at 674–76.

68. Id. at 675–77.

69. James A. Fanto, Financial Regulation Reform: Maintaining the Status Quo, 35 Brook. J. Int’l L. 635, 651–53 (2010). This Council was ultimately created by the Dodd-Frank Act and called the Financial Stability Oversight Council. See Financial Stability Oversight Council, U.S. Dep’t of Treasury, https://home.treasury.gov/policy-issues/financial-markets-financial-institutions-and-fiscal-service/fsoc [https://perma.cc/Q77W-R2ZJ] (last visited Sept. 24, 2024).

70. Cohen & Sundararajan, supra note 5, at 120, 123.

71. See Allen, supra note 5, at 1176; Brummer & Yadav, supra note 24, at 304.

72. See, e.g., Baird et al., supra note 37, at 10–12; James Bowe & William Rice, Building the Hydrogen Sector Will Require New Laws, Regs, King & Spalding (Feb. 5, 2021), https://www.kslaw.com/blog-posts/building-the-hydrogen-sector-will-require-new-laws-regs-2 [https://perma.cc/DCX4-A79A]; Damien Lyster, Vinson & Elkins, Federal Hydrogen Regulation in the United States: Where We Are and Where We Might Be Going, JDSupra (Dec. 10, 2020), https://www.jdsupra.com/legalnews/federal-hydrogen-regulation-in-the-54947/ [https://perma.cc/S5CS-3K6Y].

73. Baird et al., supra note 37, at 10, 17–22; C. Rivkin et al., Nat’l Renewable Energy Lab’y, Hydrogen Technologies Safety Guide 8, 14–26 (2015); J.L. Gillette & R.L. Kolpa, Argonne Nat’l Lab’y, Overview of Interstate Hydrogen Pipeline Systems 9–10 (2007).

74. See Baird et al., supra note 37 at 9–10, 12, 18–22 for a summary of the regulations in the hydrogen industry and the relative dearth of PHMSA and FERC regulations.

75. Coglianese & Mendelson, supra note 7, at 151–53.

76. . Cultivate Credibility, ASTM Int’l Cannabis Standards & Servs., supra note 19.

77. See Ayres & Braithwaite, supra note 7, at 101, 103, 110–11; see also Coglianese & Mendelson, supra note 7, at 161; Chavi Behl, How Do Government Regulations Impact Business Innovation and Competitiveness?, Medium (July 17, 2024), https://medium.com/@ChaviBehl/how-do-government-regulations-impact-business-innovation-and-competitiveness-6c2133c368e4 [https://perma.cc/9NSK-CFED].

78. See, e.g., Am. Petrol. Inst., supra note 34; Coglianese & Mendelson, supra note 7, at 153–54.

79. Dep’t of Com., Circular of the Bureau of Standards, No. 50: National Standard Hose Couplings and Fittings for Public Fire Service 8–10 (1914), https://nvlpubs.nist.gov/nistpubs/Legacy/circ/nbscircular50.pdf [https://perma.cc/45Y5-6AR4]; Standards, Nat’l Inst. Standards & Tech., https://www.nist.gov/standards [https://perma.cc/LN2G-5UCD] (last visited Sept. 27, 2024).

80. Momar D. Seck & David D. Evans, Fire Rsch. Div., Bldg. & Fire Rsch. Lab’y, Major U.S. Cities Using National Standard Fire Hydrants, One Century After the Great Baltimore Fire 2–5 (2004), https://www.govinfo.gov/content/pkg/GOVPUB-C13-c33f9384233e5a13eca491ede462acdf/pdf/GOVPUB-C13-c33f9384233e5a13eca491ede462acdf.pdf [https://perma.cc/XS6C-4JFQ].

81. See generally Richard L. Revesz, Rehabilitating Interstate Competition: Rethinking the “Race-to-the-Bottom” Rationale for Federal Environmental Regulation, 67 N.Y.U. L. Rev. 1210 (1992); Daniel C. Esty, Regulatory Competition in Focus, 3 J. Int’l Econ. L. 215 (2000).

82. Leiter, supra note 8, at 126, 134–35, 142–43.

83. Daniel C. Esty & Damien Geradin, Regulatory Co-opetition, 3 J. Int’l Econ. L. 235, 239–40, 243–46 (2000).

84. Id. at 242, 244; Revesz, supra note 81, at 1234, 1248. See generally Charles M. Tiebout, A Pure Theory of Local Expenditures, 64 J. Pol. Econ. 416 (1956) (setting out a theory of interjurisdictional regulatory competition).

85. Tiebout, supra note 84, at 418–19.

86. Id. at 418.

87. Richard Briffault, Our Localism: Part II—Localism and Legal Theory, 90 Colum. L. Rev. 346, 420–21 (1990) (analyzing the unrealistic nature of the mobility assumption).

88. See, e.g., Frontier Model Forum, supra note 40 (defining the criteria for membership and noting that “[t]he Forum welcomes organizations that meet these criteria”); NAESB Materials Order Form, N. Am. Energy Standards Bd., https://www.naesb.org//pdf/ordrform.pdf [https://perma.cc/W3NR-AQPY] (last updated Jan. 11, 2024) (providing an order form for standards and standard form contracts and business practices).

89. Wiegmann et al., supra note 5, at 7, 12.

90. See generally Hannah J. Wiseman & Dave Owen, Federal Laboratories of Democracy, 52 U.C. Davis L. Rev. 1119 (2018).

91. Kamilla Hvid Andersen et al., Private Standard-Setting Organizations and the Theory of Change, Bus. of Soc’y (Sept. 22, 2020), https://www.bos-cbscsr.dk/2020/09/22/the-theory-of-change/ [https://perma.cc/TCR4-2MU3].

92. Tiebout, supra note 84, at 419–20; see also Lee Anne Fennell & Richard H. McAdams, Inversion Aversion, 86 U. Chi. L. Rev. 797, 803 (2019) (describing this assumption as “highly unrealistic” in the context of local government).

93. Roberta Kwok, Kellogg: Does Positive ESG News Help a Company’s Stock Price?, MorningStar (Oct. 4, 2021), https://www.morningstar.com/insights/2021/10/04/kellogg-does-positive-esg-news-help-a-company-s-stock-price- [https://perma.cc/FT9S-KUQZ].

94. Ehud Kamar, A Regulatory Competition Theory of Indeterminacy in Corporate Law, 98 Colum. L. Rev. 1908, 1911, 1939, 1949–50 (1998); Brian Galle & Joseph Leahy, Laboratories of Democracy? Policy Innovation in Decentralized Governments, 58 Emory L.J. 1333, 1354 (2009).

95. Stein, supra note 9, at 736–37; Light, supra note 9, at 351; Joel R. Reidenberg, Governing Networks and Rule-Making in Cyberspace, 45 Emory L.J. 911, 914–16 (1996).

96. Bamigbola Paul, Emerging Industries: Understanding, Investing, and Success Stories, SuperMoney, https://www.supermoney.com/encyclopedia/emerging-industries [https://perma.cc/T2ZG-AKLY] (last updated Mar. 28, 2024).

97. Stein, supra note 9, at 747–49, 749 n.256; Light, supra note 9, at 386, 389.

98. Compare I. Glenn Cohen et al., Navigating the New Risks and Regulatory Challenges of GenAI, Harv. Bus. Rev. (Nov. 20, 2023), https://hbr.org/2023/11/navigating-the-new-risks-and-regulatory-challenges-of-genai [https://perma.cc/5D59-M75J] (discussing the need for risk regulation to prevent harms in the AI domain), with Witold J. Henisz & Bennet A. Zelner, The Hidden Risks in Emerging Markets, Harv. Bus. Rev. (Apr. 2010), https://hbr.org/2010/04/the-hidden-risks-in-emerging-markets [https://perma.cc/5JJB-X9ZN] (noting the prevalence of “policy risk,” or the “risk that a government will discriminatorily change the laws, regulations, or contracts governing an investment—or will fail to enforce them—in a way that reduces an investor’s financial returns”).

99. Legal Env’t Assistance Found., Inc. v. EPA, 118 F.3d 1467, 1469–70, 1475–76, 1478 (11th Cir. 1997).

100. Joshua P. Meltzer, The US Government Should Regulate AI if It Wants to Lead on International AI Governance, Brookings (May 22, 2023), https://www.brookings.edu/articles/the-us-government-should-regulate-ai/ [https://perma.cc/CGM4-955H]; Anthony E. DiResta & Zachary E. Sherman, The FTC Is Regulating AI: A Comprehensive Analysis, Holland & Knight (July 25, 2023), https://www.hklaw.com/en/insights/publications/2023/07/the-ftc-is-regulating-ai-a-comprehensive-analysis [https://perma.cc/JMK4-QWMJ].

101. Jeffrey M. Keisler et al., Emergent Technologies, Divergent Frames: Differences in Regulator vs. Developer Views on Innovation, Eur. J. Futures Rsch., Aug. 28, 2021, at 3, https://doi.org/10.1186/s40309-021-00180-5 [https://perma.cc/CPR8-J4R7].

102. Michael, supra note 20, at 181.

103. Bremer, supra note 14, at 324–25.

104. See Hannah Wiseman, Regulatory Adaptation in Fractured Appalachia, 21 Vill. Env’t L.J. 229, 233–34, 237–38 (2010); Leiter, supra note 8, at 126; Tara K. Righetti, Hannah J. Wiseman & James W. Coleman, The New Oil and Gas Governance, 130 Yale L.J.F. 51, 55–58, 60 (2020).

105. See Steven Mufson, How Two Small New York Towns Have Shaken Up the National Fight Over Fracking, Wash. Post (July 2, 2014, 1:14 PM), https://www.washingtonpost.com/business/economy/how-two-small-new-york-towns-have-shaken-up-the-national-fight-over-fracking/2014/07/02/fe9c728a-012b-11e4-8fd0-3a663dfa68ac_story.html [https://perma.cc/FKG5-TWPW] (noting a multitude of local governments that had banned or placed moratoria on hydraulic fracturing).

106. Leiter, supra note 8, at 128 & n.144; Wiseman & Gradijan, supra note 34, at 49.

107. Leiter, supra note 8, at 125, 128–29 (noting more than 400 local governments that had banned or placed moratoria on hydraulic fracturing).

108. John J. Koehr, Commercialization Through Standards Development, Mech. Eng’g Mag., June 2009, at 42, 42.

109. Blair Levin & Larry Downes, Who Is Going to Regulate AI?, Harv. Bus. Rev. (May 19, 2023), https://hbr.org/2023/05/who-is-going-to-regulate-ai [https://perma.cc/9VQ2-MUXF]; Frontier Model Forum, supra note 40.

110. Stein, supra note 9, at 730, 736–37.

111. See generally Leiter, supra note 8.

112. Jonathan H. Adler, Marijuana Federalism: Uncle Sam and Mary Jane 9–10, 86–89, 139 (2020).

113. Cultivate Credibility, ASTM Int’l Cannabis Standards & Servs., supra note 19.

114. See Reidenberg, supra note 95, at 914.

115. See id.; AI Watch: Global Regulatory Tracker—United States, White & Case (Dec. 18, 2024), https://www.whitecase.com/insight-our-thinking/ai-watch-global-regulatory-tracker-united-states [https://perma.cc/S4XP-8ACP]; Ryan Nabil, Artificial Intelligence Regulatory Sandboxes, 19 J.L., Econ. & Pol’y 295, 297 (2024) (noting that nations are still developing AI regulations).

116. See infra Part IV.

117. . Am. Petrol. Inst., Procedures for Standards Development 14–16 (6th ed. 2019), https://www.api.org/~/media/files/publications/2019-api-procedures-for-standards-development.pdf [https://perma.cc/ZJ9E-3SZX].

118. See generally Jacob E. Gersen, Temporary Legislation, 74 U. Chi. L. Rev. 247 (2007) (explaining the benefits of temporary legislation as it pertains to periodic sunsetting).

119. Daniel E. Walters, Lumpy Social Goods in Energy Decarbonization: Why We Need More than Just Markets for the Clean Energy Transition, 93 U. Colo. L. Rev. 541, 563–65, 578, 596 (2022). See generally Lee Anne Fennell, Slices and Lumps: Division and Aggregation in Law and Life (2019).

120. Lawrence Lessig, The Limits in Open Code: Regulatory Standards and the Future of the Net, 14 Berkeley Tech. L.J. 759, 759 (1999).

121. Id.

122. Id. at 760, 762; Rivkin et al., supra note 73, at 2737.

123. Lessig, supra note 120, at 759.

124. See Reidenberg, supra note 95, at 921, 929; infra notes 125–27 and accompanying text; Koehr, supra note 108, at 42–44.

125. See Leiter, supra note 8, at 123–24, 126.

126. See Standards, supra note 33; Leiter, supra note 8, at 148.

127. Cultivate Credibility, ASTM Int’l Cannabis Standards & Servs., supra note 19.

128. David Nevius, The History of the North American Electric Reliability Corporation 5 (2d ed. 2020), https://www.nerc.com/news/Documents/NERCHistoryBook.pdf [https://perma.cc/XN6M-GZ3E]; Macey, Welton & Wiseman, supra note 31, at 167.

129. Macey, Welton & Wiseman, supra note 31, at 179–80, 182–83.

130. . DOE Safety, Codes and Standards Activities, U.S. Dep’t of Energy, https://www.energy.gov/eere/fuelcells/doe-safety-codes-and-standards-activities [https://perma.cc/PXX7-QVEU] (last visited Sept. 11, 2024).

131. See, e.g., Wiseman, supra note 104, at 275, 284–85.

132. Coglianese & Mendelson, supra note 7, at 150. Meta-regulation has a wide variety of definitions. Under a relatively narrow version, meta-regulation “relates to corporate self-audits and safety cases where businesses develop their own rules and reporting for the regulator to assess.” F.C. Simon, Meta-Regulation in Practice: Beyond Normative Views of Morality and Rationality 2 (2017). A broader definition of meta-regulation involves formal rules that “are based on principles, not prescription” and are “reflexive and responsive.” Public law that is periodically updated based on learning, self-regulation of business within this context, and “[t]ransparency in business performance is promoted.” Id. at 2–3.

133. Brink Lindsey & Steven M. Teles, The Captured Economy: How the Powerful Enrich Themselves, Slow Down Growth, and Increase Inequality 8–10 (2017).

134. Alexis C. Madrigal, The Lies You’ve Been Told About the Origins of the QWERTY Keyboard, Atlantic, (May 3, 2013) https://www.theatlantic.com/technology/archive/2013/05/the-lies-youve-been-told-about-the-origin-of-the-qwerty-keyboard/275537/ [https://perma.cc/C873-838W].

135. Benkler, supra note 10, at 270, 272; Mansfield, supra note 10, at 173–74; Wang & Krupnick, supra note 10, at 3–6; Golden & Wiseman, supra note 10, at 961–62.

136. Liejin Guo et al., Hydrogen Safety: An Obstacle that Must Be Overcome on the Road Towards Future Hydrogen Economy, 51PD Int’l J. Hydrogen Energy 1055, 1057 (2024) (discussing safety issues); Shankari Srinivasan et al., Hydrogen: New Ambitions and Challenges, S&P Glob. (Feb. 15, 2024), https://www.spglobal.com/en/research-insights/special-reports/look-forward/hydrogen-new-ambitions-and-challenges [https://perma.cc/M8GM-X7ZJ] (discussing economic risks associated with hydrogen); see Peter Kafka, Obama: The Internet Is “The Single Biggest Threat to Our Democracy,” Vox (Nov. 16, 2020, 2:30 PM), https://www.vox.com/recode/2020/11/16/21570072/obama-internet-threat-democracy-facebook-fox-atlantic [https://perma.cc/9GB2-93KJ].

137. Statement by Vice President of Global Affairs of OpenAI, Anna Makanju, stating, “Advanced AI technologies have the potential to profoundly benefit society, and the ability to achieve this potential requires oversight and governance.” Frontier Model Forum: A New Partnership to Promote Responsible AI, Google: The Keyword (July 26, 2023), https://blog.google/outreach-initiatives/public-policy/google-microsoft-openai-anthropic-frontier-model-forum/ [https://perma.cc/LYS8-PE88]; see also supra text accompanying note 109.

138. Infrastructure Investment and Jobs Act, Pub. L. No. 117-58, 135 Stat. 429 (2021); Fact Sheet: The Bipartisan Infrastructure Deal, White House (Nov. 6, 2021), https://www.whitehouse.gov/briefing-room/statements-releases/2021/11/06/fact-sheet-the-bipartisan-infrastructure-deal/ [https://perma.cc/WKQ8-C6A2]; Inflation Reduction Act of 2022, Pub. L, No. 117-169, § 13204, 136 Stat. 1818, 1935–36 (codified at 26 U.S.C. § 45V) (creating a clean hydrogen tax credit).

139. See Steve Griffiths et al., Industrial Decarbonization Via Hydrogen: A Critical and Systematic Review of Developments, Socio-Technical Systems and Policy Options, Energy Rsch. & Soc. Sci., Oct. 2021, at 1, 9, 49 (2021).

140. Id. at 1, 1–2, 9–10, 25 (describing the “limited options” that some industrial sectors face for decarbonizing and the fact that hydrogen is one of those options).

141. James H. Williams et al., Technical and Economic Feasibility of Deep Decarbonization in the United States, in Legal Pathways to Deep Decarbonization in the United States 8, 38, 40, 66 (Michael B. Gerrard & John C. Dernbach eds., 2019).

142. See Sean O’Neil, Unlocking the Potential of Hydrogen Energy Storage, Fuel Cell & Hydrogen Energy Ass’n (July 22, 2019), https://www.fchea.org/transitions/2019/7/22/unlocking-the-potential-of-hydrogen-energy-storage [https://perma.cc/38UR-4M6C]. As with all energy sources, hydrogen is not a panacea however, using fuel as opposed to electricity to power most tasks is far less efficient and producing hydrogen fuel, in particular, is an energy-intensive endeavor. See Frank Escombe, Novel Hydroelectric Storage Concepts, in Storing Energy 67, 88 (Trevor M. Letcher ed., 2d ed. 2022) (noting that approximately four megawahours (MWh) of electricity are required to produce three MWh of zero-carbon hydrogen).

143. U.S. Dep’t of Energy, Department of Energy Hydrogen Program Plan 4 (2020), https://www.hydrogen.energy.gov/pdfs/hydrogen-program-plan-2020.pdf [https://perma.cc/5ULH-7PJ2].

144. Hydrogen Pipelines, U.S. Dep’t of Energy, https://www.energy.gov/eere/fuelcells/hydrogen-pipelines [https://perma.cc/7SWF-9TFU] (last visited Sept. 9, 2024).

145. Hydrogen Investment Pipeline Grows to $500 Billion in Response to Government Commitments to Deep Decarbonisation, Hydrogen Council (July 15, 2021), https://hydrogencouncil.com/en/hydrogen-insights-updates-july2021/ [https://perma.cc/6XM7-CUPZ]; Infrastructure Investment and Jobs Act, Pub. L. No. 117-58, § 40314, 135 Stat. 429, 1008–10 (2021).

146. Regional Clean Hydrogen Hubs Selections for Award Negotiations, U.S. Dep’t of Energy, https://www.energy.gov/oced/regional-clean-hydrogen-hubs-selections-award-negotiations [https://perma.cc/DHX7-UDFR] (last visited Sept. 10, 2024). The Department of Energy earmarked $7 billion from the IIJA to support the hubs, which will be matched by $40 billion in private investment by the selected hubs. Id.

147. Jahel Mielke & Gesine A. Steudle, Green Investment and Coordination Failure: An Investors’ Perspective, 150 Ecological Econ. 88, 93–94.

148. See Walters, supra note 119, at 548.

149. See Marcus Maher, An Analysis of Internet Standardization, Va. J.L. & Tech., Spring 1998, at 5, ¶ 15.

150. Michael Diamond, Jurisdiction over Hydrogen Pipelines and Pathways to an Effective Regulatory Regime, EBA Brief, Fall 2022, at 1, 1–3, 5–6 (noting that hydrogen gas shipped in interstate pipelines might be regulated by the Federal Energy Regulatory Commission under the Natural Gas Act, but “only if it [is] mixed with ‘natural gas,’” and that more straightforward jurisdiction over interstate hydrogen pipelines would probably require new authorizing legislation from Congress).

151. See Safe Use of Hydrogen, supra note 38.

152. Gillette & Kolpa, supra note 73, at 3–4, 9 (noting the ways in which hydrogen is safer than other fuels); cf. Miriam Ricci et al., What Do We Know About Public Perceptions and Acceptance of Hydrogen?, 33 Intl. J. Hydrogen Energy 5868, 5871 (2008) (noting that many people selected “‘No opinion’ or ‘Don’t Know’” when asked about perceptions of hydrogen safety); Roser Sala et al., Public Risk Perception of Hydrogen Technologies in Three Countries, in Advances in Reliability, Safety and Security: ESREL 2024 Collection of Extended Abstracts 185, 185 (2024) (“Public concerns about hydrogen safety and accident risk perception are key issues for the successful introduction of this technology.”).

153. See N.J. Dep’t of Health, Right to Know Hazardous Substance Fact Sheet 1–3 (2016).

154. Diamond, supra note 150, at 1, 3, 5–6.

155. See The National Fire Protection Association, NFPA, https://www.nfpa.org/ [https://perma.cc/E3D2-4PXK] (last visited Oct. 19, 2024); Fuel Cell Codes and Standards, Hydrogen Tools, https://h2tools.org/fuel-cell-codes-and-standards/nfpa-2-hydrogen-technologies-code [https://perma.cc/XW3M-PDKW] (last visited Sept. 29, 2024).

156. See Frank A. Fritz, III, Streamlining Siting, Permitting, and Construction of Hydrogen Fueling Stations Model State Statute: Updating State Fire Code Regulations 2–3 (Aug. 2020), https://lpdd.org/wp-content/uploads/2020/08/LPDD-Model-Law-State-Statute-to-Update-Fire-Code-Periodically.pdf [https://perma.cc/75HV-B49W] (noting that “[s]tates generally adopt one of the two national model fire codes: NFPA 1, Fire Code or the International Fire Code”). As Fritz further notes, many states’ adoption of these codes leaves out critical hydrogen standards that have been promulgated by NFPA more recently. Id. at 3.

157. See infra notes 162–67 and accompanying text.

158. See The History of the NFPA, Inspect Point (July 7, 2019), https://www.inspectpoint.com/history-of-the-nfpa/ [https://perma.cc/W7C3-T6DM]; Casey Cavanaugh Grant, History of NFPA, NFPA, https://web.archive.org/web/20231017084348/https:/www.nfpa.org/About-NFPA/NFPA-overview/History-of-NFPA [https://perma.cc/TAK8-VBP3] (last visited Jan. 2, 2025).

159. See The History of the NFPA, supra note 158.

160. Underwriters Laboratories Inc., Engineering Progress: The Revolution and Evolution of Working for a Safer World 5, 7–10 (2016).

161. Id. at 10–11, 15.

162. The History of the NFPA, supra note 158.

163. Id.

164. Id.

165. Dominic Cuthbert, How the Hindenburg Haunted Hydrogen (and Why It Needs Laying to Rest), ITM Power (Aug. 1, 2022), https://itm-power.com/blogs/how-the-hindenburg-haunted-hydrogen [https://perma.cc/5KWE-QEVA].

166. U.S. Dep’t of Energy, Fact Sheet Series: Hydrogen Safety 1, 2, https://www1.eere.energy.gov/hydrogenandfuelcells/pdfs/h2_safety_fsheet.pdf [https://perma.cc/LB74-XF7M] (last visited Sept. 11, 2024).

167. Id.; see Fuel Cell Codes and Standards, supra note 155.

168. NFPA 2: Hydrogen Technologies Code, NFPA, https://www.nfpa.org/codes-and-standards/2/2 [https://perma.cc/BR89-HUKQ] (last visited Jan. 2, 2025).

169. NFPA Makes Important Safety Codes and Standards Available for Free Online, NFPA, https://www.nfpa.org/For-Professionals/Codes-and-Standards/List-of-Codes-and-Standards/Free-access [https://perma.cc/HRQ4-4A95] (last visited Sept. 11, 2024).

170. A.P. Harris et al., Sandia Nat’l Lab’ys, Safety, Codes and Standards for Hydrogen Installations: Metrics Development and Benchmarking 9–10 (2014), https://www.osti.gov/servlets/purl/1177046 [https://perma.cc/EF6S-REWM].

171. Hydrogen Fuel Cell P’ship, NFPA 2 and the California Fire Code (2022), https://h2fcp.org/sites/default/files/NFPA-2-and-California-Fire-Code.pdf [https://perma.cc/J3Z4-62Z5].

172. Rivkin et al., supra note 73, at 13; see Key Changes to NFPA 25 in the 2023 Edition: What You Need to Know, Inspect Point (Sept. 12, 2024), https://www.inspectpoint.com/changes-to-nfpa-25-2023-edition/ [https://perma.cc/ZPT8-TXBL].

173. See infra notes 186–87 and accompanying text.

174. See supra Section IV.A.

175. Joseph Romm, California’s Hydrogen Highway Reconsidered, 36 Golden Gate U. L. Rev. 393, 405 (2006); Air Res. Bd., Cal. Env’t Prot. Agency, California Exhaust Emission Standards and Test Procedures for 2005 Through 2008 Model Zero-Emission Vehicles, and 2001 Through 2008 Model Hybrid Electric Vehicles, in the Passenger Car, Light-Duty Truck and Medium-Duty Vehicle Classes at A, B-1, C-1–C-2 (2009), https://ww2.arb.ca.gov/sites/default/files/2020-07/clean_2005_2008_my_hev_tps_12_09_ac.pdf [https://perma.cc/W4EC-L8LM].

176. Romm, supra note 175, at 405. It is possible, though less desirable, to burn hydrogen in modified internal combustion engines. Id. at 402.

177. Evan Halper, Is California’s ‘Hydrogen Highway’ a Road to Nowhere?, L.A. Times (Aug. 10, 2021, 3:00 AM), https://www.latimes.com/politics/story/2021-08-10/hydrogen-highway-or-highway-to-nowhere [https://perma.cc/7UEX-SN32].

178. See id.

179. Cal. State Fire Marshal Info. Bull. 14-010, Adoption of NFPA 2 Hydrogen Technologies Code for the Supplement to the 2013 California Building and Fire Code (2014), https://34c031f8-c9fd-4018-8c5a-4159cdff6b0d-cdn-endpoint.azureedge.net/-/media/osfm-website/resources/information-bulletins/ib_14010codesupplementnfpa2.pdf [https://perma.cc/RV4W-GSD5].

180. Halper, supra note 177.

181. Id.

182. NREL Learning, Permitting Hydrogen Fueling Stations, YouTube (2016), https://www.youtube.com/playlist?list=PLmIn8Hncs7bE9xYhaYKj9kDnE-hNy5eTe [https://perma.cc/8Z64-FK22].

183. Halper, supra note 177.

184. About Us, Fuel Cell & Hydrogen Energy Ass’n, https://fchea.org/about-fchea/ [https://perma.cc/7DUH-X8AD] (last visited Sept. 21, 2024); see Credits for New Clean Vehicles Purchased in 2023 or After, Internal Revenue Serv., https://www.irs.gov/credits-deductions/credits-for-new-clean-vehicles-purchased-in-2023-or-after [https://perma.cc/Z3Y8-QJL9] (last updated Aug. 8, 2024).

185. Tatarenko et al., supra note 35.

186. Carl Rivkin et al., A National Set of Hydrogen Codes and Standards for the United States, 36 Int’l J. Hydrogen Energy 2736, 2737 (2011).

187. See generally Susan Bershad, Nat’l Fire Prot. Ass’n, NFPA 2 Overview: SoCal Fire Prevention Officers 3, 30 (2018), https://www.ourair.org/wp-content/uploads/THRPAppendixC7.pdf [https://perma.cc/C7RL-VX3W].

188. About Us, Fuel Cell & Hydrogen Energy Ass’n, supra note 184.

189. Regulations, Codes and Standards, Fuel Cell & Hydrogen Energy Ass’n, https://fchea.org/our-work/regulations-codes-and-standards/ [https://perma.cc/T3YM-YWJN] (last visited Sept. 21, 2024).

190. Am. Inst. of Aeronautics & Astronautics & Am. Nat’l Standards Inst., Guide to Safety of Hydrogen and Hydrogen Systems 1 (2017), https://webstore.ansi.org/preview-pages/AIAA/preview_ANSI+AIAA+G-095A-2017.pdf?srsltid=AfmBOoqeiBFCbmpUf-yG9I1PPrieedqpa5qzB__M5p3i4CFcBzZZ9jfy [https://perma.cc/T28F-8JYY].

191. Safety, Codes, and Standards, NREL https://www.nrel.gov/hydrogen/safety-codes-standards.html [https://perma.cc/E2WK-5CWC] (last visited Sept. 24, 2024); Mattie Hensley, Improving Hydrogen Safety Codes and Standards, Sandia LabNews (Aug. 25, 2022), https://www.sandia.gov/labnews/2022/08/25/improving-hydrogen-safety-codes-and-standards/ [https://perma.cc/VX78-RV4G]; Harris et al., supra note 170, at 19.

192. Baird et al., supra note 37, at 7.

193. See Fuel Cell Codes and Standards: Comprehensive Guide, Hydrogen Tools, https://h2tools.org/fuel-cell-codes-and-standards [https://perma.cc/6HCF-2GYQ] (last visited Dec. 28, 2024).

194. Ashley Gibson & Matt Scoular, The Global Hydrogen Economy in 2023—Ditching the Colours in Favour of Quantifiable Standards, SLR Consulting (Aug. 22, 2023), https://www.slrconsulting.com/insights/the-global-hydrogen-economy-in-2023-ditching-the-colours-in-favour-of-quantifiable-standards [https://perma.cc/HQD3-BMH4].

195. See Rivkin et al., supra note 186, at 2738–40.

196. See Rivkin et al., supra note 73, at 12, 34.

197. Baird et al., supra note 37, at 18–21.

198. Koehr, supra note 108, at 42, 44.

199. CGA PS-48 CGA Position Statement on Clarification of Existing Hydrogen Setback Distances and Development of New Hydrogen Setback Distances in NFPA 55, Hydrogen Tools, https://h2tools.org/fuel-cell-codes-and-standards/cga-ps-48-cga-position-statement-clarification-existing-hydrogen [https://perma.cc/262H-ABCX] (last visited Jan. 2, 2025).

200. Id.

201. Chris LaFleur, Sandia Nat’l Lab’ys, Gaseous Hydrogen Separation Distances (2017), https://www.osti.gov/servlets/purl/1429299 [https://perma.cc/FM85-A2LL].

202. Cf. Romm, supra note 175, at 400, 405, 407 (noting the “chicken and egg” problem for hydrogen fueling infrastructure).

203. See Lyster, supra note 72.

204. See Safety, Codes and Standards Basics, U.S. Dep’t of Energy, https://www.energy.gov/eere/fuelcells/safety-codes-and-standards-basics [https://perma.cc/4RHE-2QJP] (last visited Sept. 22, 2024).

205. How Are Building Codes Adopted? U.S. Dep’t of Energy (Sept. 19, 2016), https://www.energy.gov/eere/buildings/articles/how-are-building-codes-adopted [https://perma.cc/3Z77-HQSY].

206. See Codes and Standards, U.S. Dep’t of Energy, https://www.energy.gov/eere/fuelcells/codes-and-standards [https://perma.cc/X4XP-97RS] (last visited Jan. 2, 2025) (stating “[e]fforts are underway to support activities at the International Code Council and the National Fire Protection Association for the development and adoption of their family of model codes” and later addressing hydrogen as a key code).

207. Codes and Standards Activities, U.S. Dep’t of Energy, https://www.energy.gov/eere/fuelcells/codes-and-standards-activities [https://perma.cc/N24R-QQKM] (last visited Sept. 22, 2024).

208. See generally Carl Rivkin et al., Nat’l Renewable Energy Lab’y, Guide to Permitting Hydrogen Motor Fuel Dispensing Facilities (2016), https://h2tools.org/sites/default/files/Motor Fueling Station Permit Guide Final March2016.pdf [https://perma.cc/W53Z-SP6N]; H2IQ Hour Webinars, U.S. Dep’t of Energy, https://www.energy.gov/eere/fuelcells/hydrogen-and-fuel-cell-technologies-office-webinars [https://perma.cc/EE49-Y5GF] (last visited Sept. 29, 2024); Codes & Standards: Permitting Tools, Hydrogen Tools, https://h2tools.org/codes-standards/codes-standards-permitting-tools [https://perma.cc/RCY5-Q6XS] (last visited Sept. 29, 2024).

209. . Regulations, Guidelines, and Codes and Standards, U.S. Dep’t of Energy, https://www.energy.gov/eere/fuelcells/regulations-guidelines-and-codes-and-standards [https://perma.cc/NQD3-4Y8N] (last visited Sept. 9, 2024); Hydrogen Program: Codes and Standards, U.S. Dep’t of Energy, https://www.hydrogen.energy.gov/program-areas/codes-standards [https://perma.cc/Y3S2-SXZS] (last visited Sept. 22, 2024); Fuel Cell Codes and Standards: Comprehensive Guide, supra note 193.

210. See supra Section IV.A.

211. Koehr, supra note 108, at 42.

212. Mike Soraghan, Biden Energy Agenda Exposes Regulatory Gap, E&E News (Feb. 6, 2023, 6:56 AM) https://www.eenews.net/articles/biden-energy-agenda-exposes-regulatory-gap/ [https://perma.cc/AJE4-4RLK]; Christopher Psihoules & Daniel Salomon, Hydrogen Pipeline Regulation, Norton Rose Fulbright (June 23, 2023), https://www.projectfinance.law/publications/2023/june/hydrogen-pipeline-regulation/ [https://perma.cc/RLU2-VHW9].

213. See, e.g., Carl H. Rivkin, Nat’l Renewable Energy Lab’y, A National Set of Hydrogen Codes and Standards for the US (2009), https://www.nrel.gov/docs/fy09osti/46604.pdf [https://perma.cc/FTV7-ZRDE] (noting “[h]ydrogen dispensing requirements added to NFPA 52 Code”).

214. See Stephanie Wong, Note, Self-Regulation by the Private Industry and Its Effectiveness in Today’s Online Environment, J. on Emerging Techs., Feb. 2022, at 161, 163–64.

215. Daniel Benoliel, Cyberspace Technological Standardization: An Institutional Theory Retrospective, 18 Berkeley Tech. L.J. 1259, 1285 (2003); Maher, supra note 149, at ¶ 4.

216. Llewellyn J. Gibbons, No Regulation, Government Regulation, or Self-Regulation: Social Enforcement or Social Contracting for Governance in Cyberspace, 6 Cornell J.L. & Pub. Pol’y 475, 481, 492, 501, 509 (1997).

217. William J. Clinton & Albert Gore, Jr., A Framework for Global Electronic Commerce 4 (1997); Edward Wyatt, Obama Asks F.C.C. to Adopt Tough Net Neutrality Rules, N.Y. Times (Nov. 10, 2014), https://www.nytimes.com/2014/11/11/technology/obama-net-neutrality-fcc.html [https://perma.cc/DUQ3-WMG8].

218. Wyatt, supra note 217.

219. See Lessig, supra note 120, at 762–63; Neil Weinstock Netanel, Cyberspace Self-Governance: A Skeptical View from Liberal Democratic Theory, 88 Calif. L. Rev. 395, 473, 476, 479 (2000).

220. David R. Johnson & David Post, Law and Borders—The Rise of Law in Cyberspace, 48 Stan. L. Rev. 1367, 1367, 1390, 1391 (1996).

221. A. Michael Froomkin, Habermas@discourse.net: Toward of Critical Theory of Cyberspace, 116 Harv. L. Rev. 749, 783 (2003).

222. Id. at 784, 786–87.

223. Id. at 786–87.

224. Justus Baron et al., Balance and Standardization: Implications for Competition and Antitrust Analysis, 84 Antitrust L.J. 425, 440 (2022).

225. Id. at 449.

226. Id.

227. Froomkin, supra note 221, at 778–79.

228. Id. at 787.

229. Id. at 820–21 (footnote omitted).

230. Id. at 799, 819–20.

231. See, e.g., Esma Aïmeur et al., Fake News, Disinformation and Misinformation in Social Media: A Review, Soc. Network analysis & Mining, Feb. 9, 2023, Art. No. 13, at 1, 12 (2023).

232. Philip J. Weiser, The Internet, Innovation, and Intellectual Property Policy, 103 Colum. L. Rev. 534, 541–42 (2003).

233. Netanel, supra note 219, at 399–400 (footnotes omitted).

234. Baron et al., supra note 224, at 440.

235. Maher, supra note 149, at ¶ 15.

236. Id. at ¶ 78.

237. Paul W. Parfomak, Cong. Rsch. Serv., R46700, Pipeline Transportation of Hydrogen: Regulation, Research, and Policy 9 (2021).

238. Gibbons, supra note 216, at 502.

239. Am. Librs. Ass’n v. Pataki, 969 F. Supp. 160, 164 (S.D.N.Y. 1997).

240. Id. at 177.

241. NIST, Artificial Intelligence Risk Management Framework (AI RMF 1.0) 1 (2023), https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf [https://perma.cc/TVF9-29QS] (adapted from the OECD and ISO/IEC definitions).

242. Meltzer, supra note 100.

243. Stan. Univ., What Are the Most Important Advances in AI?, in Gathering Strength, Gathering Storms: The One Hundred Year Study on Artificial Intelligence (AI100) 2021 Study Panel Report 12, 16–17 (2021), https://ai100.stanford.edu/sites/g/files/sbiybj18871/files/media/file/AI100Report_MT_10.pdf [https://perma.cc/NYD2-SN3G].

244. Id. at 12, 17.

245. See, e.g., Danielle Abril, AI Isn’t Yet Going to Take Your Job—But You May Have to Work with It, Wash. Post (Mar. 20, 2023, 9:22 AM), https://www.washingtonpost.com/technology/interactive/2023/ai-jobs-workplace/ [https://perma.cc/KJJ4-ASLX]; Gretchen Tarrant, Generative AI Is Already Changing White-Collar Work as We Know It, Wall St. J. (Mar. 29, 2023, 10:00 AM), https://www.wsj.com/articles/generative-ai-is-already-changing-white-collar-work-as-we-know-it-58b53918 [https://perma.cc/A6HC-88S5].

246. But see Tejas N. Narechania, Machine Learning as Natural Monopoly, 107 Iowa L. Rev. 1543, 1545 (2022) (arguing that AI does involve networks that function as a natural monopoly and suggesting that it should therefore be tightly regulated as a traditional utility).

247. See GPT Actions, OpenAI, https://platform.openai.com/docs/plugins/introduction [https://perma.cc/ARZ6-57RA] (last visited Sept. 7, 2024); Aaron Hurst, Microsoft Extends Generative AI for Suite and Third-Party Software, Info. Age (May 24, 2023), https://www.information-age.com/microsoft-extends-generative-ai-for-suite-third-party-software-123504149/ [https://perma.cc/JY98-PKEK].

248. See Narechania, supra note 246, at 1573. Despite public fears about potential harms from AI, it is marching forward at a rapid pace. Lee Rainie et al., How Americans Think About Artificial Intelligence, Pew Rsch. Ctr. (Mar. 17, 2022), https://www.pewresearch.org/internet/2022/03/17/how-americans-think-about-artificial-intelligence/ [https://perma.cc/23V4-8YRM] (“In broad strokes, a larger share of Americans say they are ‘more concerned than excited’ by the increased use of AI in daily life than say the opposite.”). Contrast this with hydrogen, which also inspires public concern and has not yet been commercialized to a meaningful extent. See supra text accompanying note 108.

249. Joshua A. Goland, Algorithmic Disgorgement: Destruction of Artificial Intelligence Models as the FTC’s Newest Enforcement Tool for Bad Data, 29 Rich. J.L. & Tech. 1, 3, (2023) (quoting Ashkan Soltani, former FTC Chief Technologist, now an “independent researcher and technologist”); About, Ashkan Soltani https://ashkansoltani.org/ [https://perma.cc/76YU-NPKL].

250. Allen, supra note 5, at 1176.

251. Id.

252. Stan. Univ., What Are the Most Pressing Dangers of AI?, in Gathering Strength, Gathering Storms: The One Hundred Year Study on Artificial Intelligence (AI100) 2021 Study Panel Report, supra note 243, at 53, 53–56.

253. DiResta & Sherman, supra note 100.

254. Id.

255. See Goland, supra note 249, at 12; DiResta & Sherman, supra note 100 (footnote omitted).

256. Jess Weatherbed, Open AI’s Regulatory Troubles Are Only Just Beginning, Verge (May 5, 2023, 5:30 AM), https://www.theverge.com/2023/5/5/23709833/openai-chatgpt-gdpr-ai-regulation-europe-eu-italy [https://perma.cc/X24Z-ZUTV]; Intelligenza Artificiale: il Garante Blocca Chat GPT. Raccolta Illecita di Dati Personali. Assenza di Sistemi per la Verifica Dell’eta dei Minori, Garanta per la Protezione dei Dati Personali (Mar. 31, 2023), https://www.garanteprivacy.it/home/docweb/-/docweb-display/docweb/9870847# [https://perma.cc/J4Q5-ULHZ].

257. Lauren Feiner, OpenAI Faces Complaint to FTC that Seeks Investigation and Suspension of ChatGPT Releases, CNBC, https://www.cnbc.com/2023/03/30/openai-faces-complaint-to-ftc-that-seeks-suspension-of-chatgpt-updates.html [https://perma.cc/FUQ4-WSB9] (last updated Apr. 17, 2023, 1:48 AM).

258. Frontier Model Forum, supra note 40; Progress Update: Advancing Frontier AI Safety in 2024 and Beyond, Frontier Model F. (Aug. 29, 2024), https://www.frontiermodelforum.org/updates/progress-update-advancing-frontier-ai-safety-in-2024-and-beyond/ [https://perma.cc/PZD5-VX8X].

259. Usage Policies, OpenAI API, http://120.25.75.130:10033/doc/172 [https://perma.cc/S8DP-W8JC] (last updated Mar. 23, 2023).

260. Id.

261. Id.

262. AI Risk Management Framework, NIST, https://www.nist.gov/itl/ai-risk-management-framework [https://perma.cc/59TV-AKFT] (last visited Sept. 26, 2024).

263. Id.

264. NIST, supra note 241, at 2–3.

265. Id. at 12–17.

266. Id. at 16.

267. Russell T. Vought, Exec. Off. of the President, M-21-06, Guidance for Regulation of Artificial Intelligence Applications 1–2 (Nov. 17, 2020), https://www.whitehouse.gov/wp-content/uploads/2020/11/M-21-06.pdf [https://perma.cc/BE69-LZ7E]

268. Id. at 2.

269. Id. at 3–7.

270. Shalanda D. Young, Exec. Off. of the President, M-24-10, Advancing Governance, Innovation, and Risk Management for Agency Use of Artificial Intelligence 3 n.7 (Mar. 28, 2024), https://www.whitehouse.gov/wp-content/uploads/2024/03/M-24-10-Advancing-Governance-Innovation-and-Risk-Management-for-Agency-Use-of-Artificial-Intelligence.pdf [https://perma.cc/XBN5-C8BP] (citing AI in Government Act of 2020 and Executive Order 13960).

271. Allen, supra note 5, at 1175 (citing to the Clinical Laboratory Improvement Amendments of 1988).

272. The White House Launches the National Artificial Intelligence Initiative Office, Off. of Sci. & Tech. Pol’y (Jan. 12, 2021), https://trumpwhitehouse.archives.gov/briefings-statements/white-house-launches-national-artificial-intelligence-initiative-office/ [https://perma.cc/93X8-QGK6]; NSF Advances Artificial Intelligence Research with New Nationwide Institutes, Nat’l Sci. Found. (Aug. 26, 2020), https://www.nsf.gov/news/special_reports/announcements/082620.jsp [https://perma.cc/76FH-PMFC].

273. See supra text accompanying notes 206–10.

274. See supra text accompanying note 260.

275. FACT SHEET: Vice President Harris Announces OMB Policy to Advance Governance, Innovation, and Risk Management in Federal Agencies’ Use of Artificial Intelligence, White House (Mar. 28, 2024), https://www.whitehouse.gov/briefing-room/statements-releases/2024/03/28/fact-sheet-vice-president-harris-announces-omb-policy-to-advance-governance-innovation-and-risk-management-in-federal-agencies-use-of-artificial-intelligence/# [https://perma.cc/RL3C-NSSV].

276. Meltzer, supra note 100.

277. Written Testimony of Sam Altman, Chief Executive Officer, OpenAI, Before the Senate Committee on the Judiciary Subcommittee on Privacy, Technology, & the Law (2023), https://www.judiciary.senate.gov/imo/media/doc/2023-05-16 - Bio & Testimony - Altman.pdf [https://perma.cc/5WH7-HACN].

278. Id.

279. Meltzer, supra note 100 (describing and contrasting the testimony).

280. Id.; The Forum for Cooperation on Artificial Intelligence, Brookings, https://www.brookings.edu/projects/the-forum-for-cooperation-on-artificial-intelligence/ [https://perma.cc/W84P-J8JW] (last visited Jan. 2, 2025).

281. See supra note 108 and accompanying text.

282. Maher, supra note 149, at ¶ 14.

283. See supra text accompanying notes 185–87; Koehr, supra note 108, at 43–44; NGN Secures Funding to Explore Hydrogen Potential, Transport & Energy (Jan. 17, 2024), https://transportandenergy.com/2024/01/17/ngn-secures-funding-to-explore-hydrogen-potential/ [https://perma.cc/2RZ7-2WHE].

284. Glossary: Cyberspace, NIST, https://csrc.nist.gov/glossary/term/cyberspace [https://perma.cc/6AHN-RPYQ] (last visited Sept. 27, 2024).

285. Lessig, supra note 120, at 759–60.

286. Id. at 759.