Regulatory Comments

ICLE Comments to ACCC’s Digital Platform Services Inquiry

Introduction

We thank the Australian Competition & Consumer Commission (ACCC) for the invitation to comment on its July 25 issues paper on recent developments and emerging issues in digital-platform markets, in anticipation of the 10th and final report of its Digital Platform Services Inquiry.[1] The International Center for Law & Economics (ICLE) is a nonprofit, nonpartisan global research and policy center founded with the goal of building the intellectual foundations for sensible, economically grounded policy. ICLE promotes the use of law & economics methodologies to inform public-policy debates and has longstanding expertise in the evaluation of competition law and policy. ICLE’s interest is to ensure that competition law remains grounded in clear rules, established precedent, a record of evidence, and sound economic analysis.

We commend the ACCC’s cautious approach to intervening in digital markets by conducting a comprehensive inquiry before suggesting digital market regulations or reforms to competition law to address alleged competition problems in those markets. As highlighted by the ACCC chair in the press release announcing this paper, “in many cases international legislation is only recently enacted so the full impact may not be clear.”[2]

In these comments, we explore, from a competition and regulation perspective, some of the issues addressed in the issues paper. We respectfully suggest careful consideration before approving any sectoral regulation of digital markets in Australia. Digital markets are generally dynamic, competitive, and beneficial to consumers. Those benefits derive mainly from increased productivity and relatively cheap access to information. While there are always possible competition issues and anticompetitive behavior, these are neither pervasive nor sufficiently unique to justify strict, sui generis preemptive rules. We acknowledge that the ACCC:

[A]grees with the growing international consensus that digital platforms require specific and tailored regulation. While various jurisdictions are taking different approaches to implementing such measures, it is clear to the ACCC that enforcing existing competition and consumer laws ‘ex post’ (i.e. after conduct has occurred) cannot by itself address the systemic and significant problems arising in markets for digital platform services.[3]

We respectfully disagree and posit, instead, that existing antitrust laws are sufficient to address potential anticompetitive practices in digital markets, as we explain below. Furthermore, we think that the ACCC already has the necessary tools and expertise to handle these cases.

Of course, applying antitrust laws to digital markets can be challenging. For example, it can be difficult to define relevant markets and dominant positions on multisided platforms and in the fast-changing digital landscape. The identity of relevant competitors and competing products and services is not always clear, and the boundaries between the digital and non-digital worlds are sometimes overstated. Those challenges, however, can be properly addressed through the existing legal framework. Moreover, many limitations on the effectiveness of ex-post enforcement can be overcome with institutional measures, such as equipping the ACCC with more resources to incorporate advanced, state-of-the-art technical expertise.

Ex-ante regulations like the European Union’s Digital Markets Act (DMA) can have serious unintended consequences, such as stifling innovation, reducing consumer welfare, and increasing compliance costs. They can also lead to increased risks of regulatory capture and rent-seeking, as the verdict on whether a gatekeeper has complied with the law often comes down to the degree to which rivals are satisfied. Of course, rivals have a clear personal stake in never being satisfied. By tethering intervention to a comparatively clear public-benefit standard—consumer welfare—competition laws minimize the potential for error costs and decrease the chances that the law will be coopted for private gain.

Our comments also express the view that policymakers’ present concerns about competition in “AI markets” may be unwarranted. This is particularly true of the notions that data-driven network effects shield incumbents in AI markets from competition; that Web 2.0’s most successful platforms leverage their competitive positions to dominate generative-AI markets; that these same platforms use strategic partnerships with AI firms to insulate themselves from competition; and that generative-AI services occupy narrow markets that leave firms with significant market power.

In fact, we are still far from understanding the boundaries of antitrust-relevant markets in AI. Three primary notions should be at the forefront of competition authorities’ minds when they think about market definitions surrounding AI products and services. First, the “AI market” is not unitary, but instead comprises many distinct goods and services. Second, and relatedly, despite AI marketing hype, this extremely heterogeneous product landscape intersects with equally variegated consumer demand.

In other words: AI products and services may, in many instances, be substitutable for non-AI products, which would mean that, for the purposes of antitrust law, AI and non-AI products contend in the same relevant markets. Getting this relevant product-market definition right is important in antitrust, because incorrect market definitions could lead to wrong inferences about market power. While either an overly broad or overly narrow market definition could lead to erroneous enforcement, we believe the former currently represents the bigger threat.

Third, overenforcement in the field of generative AI could paradoxically engender the very harms that policymakers are seeking to avert. As we explain in greater detail below, preventing so-called “Big Tech” firms from competing in AI markets (for example, by threatening competition intervention whenever they forge strategic relationships with AI startups, launch their own generative-AI services, or embed such services in their existing platforms) may thwart an important source of competition and continued innovation. Competition in AI markets is important,^[4] but trying naïvely to hold incumbent tech firms back out of misguided fears they will come to dominate the AI space is likely to do more harm than good. It is essential to acknowledge how little we know about these nascent markets and that the most important priority at the moment is to ask the right questions to ensure sound competition policy.

The remainder of these comments proceeds as follows: Section II considers international regulatory developments and major market developments discussed in the issues paper’s Topics 1 and 2. Section III considers the emerging issue of competition and AI, aiming to provide a balanced view in a discussion in which proponents of “preemptive enforcement” may be amplifying potential competitive threats that are currently speculative.

II. International Regulatory Developments and Major Developments in Digital Platforms: Rationales for Digital Competition Regulation Are Uncertain

The issues paper presents a detailed description of recent regulatory developments (including both the enactment of new regulations and the discussion of bills and expert reports) in the European Union, Germany, India, Japan, South Korea, and the United Kingdom. As it notes, some key themes have emerged, including “addressing anti-competitive self-preferencing, tying and bundling practices, addressing barriers to switching, and promoting access to third-party applications on a platform or service”.[5]

Whether these are explicit regulations (such as the Digital Markets Act (DMA) in the European Union) or reforms to competition law (as in Germany), all of these reforms seek to ban some or all of the aforementioned conduct or business models due to their allegedly negative effects on competition.

In general terms, “digital markets” or digital platforms are not good candidates for ex-ante regulation, given that the markets for such services are reasonably competitive. As Herbert Hovenkamp has explained:

With Big Tech, we’re looking at probably the most productive part of the economy. The rate of innovation is high. They spend a lot of money on R&D. They are among the largest patent holders. There’s very little evidence of collusion. They seem to be competing with each other quite strongly. They pay their workers relatively well and have fairly educated workforces. None of this is a sign that these are industries we should be pursuing. That doesn’t mean they don’t do some anti-competitive things. But the whole idea that we should be targeting Big Tech strikes me as fundamentally wrong-headed.[6]

According to economic theory and long-tested economic principles, ex-ante regulation^[7] is justified only in the presence of clear market failures^[8]. Digital markets, however, do not present the kind of market failures that warrant ex-ante regulation. For example, digital markets do not generally foment natural monopolies, significant externalities, public goods, or informational asymmetries.

To be sure, one can find some level of informational asymmetries or externalities in some digital markets, but not of such magnitude that they couldn’t be addressed through market competition (actual or potential) or through general rules, such as data-protection or consumer-protection laws. A more plausible argument can be made regarding the presence of “network effects” in online platforms. If a firm moves fast and is the first to attract customers, that customer base will, in turn, attract more customers and sellers. This network growth could, so the story goes, result in a single firm monopolizing the market. However, as Evans and Schmalensee have pointed out, that result is far from inevitable:

Systematic research on online platforms by several authors, including one of us, shows considerable churn in leadership for online platforms over periods shorter than a decade. Then there is the collection of dead or withered platforms that dot this sector, including Blackberry and Windows in smartphone operating systems, AOL in messaging, Orkut in social networking, and Yahoo in mass online media.^[9]

Some regulations and proposals such as the DMA or the proposed American Innovation and Choice Online Act (AICOA) in the United States mention alleged failures of antitrust law (i.e., “too slow” and “too hard for plaintiffs”) as the primary rationale to regulate digital markets. As Giuseppe Colangelo has explained:

Against this background, the regulatory approaches recently advanced do not seem to reflect the distinctive features of digital markets, but rather the need to design enforcement short-cuts to cope with growing concerns that antitrust law is unable to address potential anticompetitive practices by large online platforms. Hence, in most of the mentioned reports, the revival of regulation seems supported more by an alleged antitrust enforcement failure rather than true a market failure. The goal is indeed to fill alleged enforcement gaps in the current antitrust rules by introducing tools aimed at lowering legal standards and evidentiary burdens in order to address anti-competitive practices that standard antitrust analysis would struggle to tackle.^[10]

The ACCC’s fifth interim report does not rely on the presence of market failures to support the need for new competition measures and/or regulations. Rather, resembling the discussion about the above-mentioned regulations and proposals in the European Union and the United States, the ACCC considers that “the enforcement of traditional competition laws against digital platforms in Australia would likewise generally be a slow process. Further, the immense scale and financial resources of large digital platforms may impede traditional enforcement through the courts, resulting in protracted litigated outcomes.”[11]

Sluggish procedure could be a plausible justification for piecemeal reform. The ACCC’s fifth interim report acknowledges this, giving several examples of recent antitrust cases in both Australia and Europe.[12] Antitrust cases could possibly be better expedited. Competition agencies and courts should generally have more resources and faster procedures to adjudicate cases before market structures or markets in general change, rendering any potential intervention useless. This can, however, be resolved via institutional and legal reforms within existing competition law.

The fact that cases are deemed “hard to win” is not a valid justification to jettison the law. Indeed, this might be a strength, not a shortcoming, of antitrust law, particularly in the context of “abuse of dominance” or monopolization cases.^[13] Antitrust cases start from the premise that it is not always immediately obvious whether certain conduct is pro- or anticompetitive.[14] What follows is a structured inquiry operating within a burden-shifting framework that aims to elucidate this question. Regulations like the DMA jettison this cautious approach by condemning broad swathes of specific conduct with strict ex-ante rules that cover many different companies, in different sectors, and with a range of varied products and services. Competitive conditions vary widely, as do the competitive consequences of the business models employed by these firms. DMA-style regulations substitute theoretically robust concepts like “relevant markets” and “market power” or “dominant position” for cruder ones like “core platforms services” or “gatekeeper,” with the express intent of providing shortcuts to condemn business models and practices. But these “shortcuts” have a cost: they can easily lead to condemnation of business models and practices that provide benefits for consumers, such as lower prices and a safer user experience, among others.

Even those open to considering digital-markets regulation acknowledge that there are considerable challenges, especially if the intent is to regulate digital platforms in a manner similar to “essential facilities”:

In the tech industry, the first challenge is to identify a stable essential facility. It must be stable because divestitures take a while to perform, and the cost of implementing them would not be worth its while if the location of the essential facility kept migrating. This condition may not be met, though. While the technology and market segments of electricity, railroads, and (up to the 1980s) telecoms had not changed much since the early twentieth century, digital markets are fastmoving. Rapidly morphing technologies and demands make it difficult for regulators to identify, collect data on, and regulate essential facilities.^[15]

Moreover, even if warranted, regulations create barriers to entry and regulatory risks, and they restrict the monetization of business assets. They also tend to make markets less attractive and could deter potential competitors from entering them. It is possible that the DMA is already producing such consequences. As Alba Ribera has explained:

One of the greatest examples of the dichotomy that arises between the different types of consequences that can be generated by the regulatory capture of digital ecosystems can be found in Meta’s recent decision not to launch its new service Threads in the European Economic Space. To the extent that its service could be interpreted as falling within the definition of a “core platform service” belonging to the category of “online social networks” (listed by the DMA), Meta decided to refrain from entering the European market, due to the disproportionate burden that the demanding obligations imposed by the DMA would entail. It should be noted that Threads is still an entrant service in the online social networking market, in contrast to the predominant position occupied by X (previously known as Twitter). In this way, we observe that the categorization as a core platform service unifies and eliminates all the nuances that free competition entails with respect to incoming services in the markets.^[16]

Some of these unintended consequences were observable in the EU even before the DMA fully entered into force. From users’ perspective, regulation can serve to make services and products more expensive. Facebook is already experimenting with a new business model in the EU in which the consumer would see no ads (thus, there would be no data collection, or less collection of data for marketing purposes, at any rate), but would have to pay for subscriptions.[17] If this business model would generalize, some privacy-minded users may prefer it and probably would be able to afford it. But other consumers that may prefer and have benefitted the most from digital platforms with zero price or otherwise affordable products, such as WhatsApp and Facebook, would be worse off.

From the perspective of the companies that own and operate digital platforms and services, if regulations like the DMA make their platforms less profitable, some could choose not to enter or, indeed, to leave such markets. As Geoffrey Manne and Dirk Auer have explained, “to regulate competition, you first need to attract competition.”^[18]

While these considerations are especially pertinent in the context of developing countries, which rely heavily on attracting foreign direct investment, they could also affect the competitiveness of countries like Australia.

The DMA entered into effect in full force in March 2024. While it may be too early to reach definitive conclusions about its effects, consumers are already reporting that they have experienced a degraded user experience. For example, the French newspaper Liberation has detailed how Google Maps’ map results are not showing directly in search-results pages in the same ways they once did (See Figures 1 and 2).

FIGURE 1: US Search Results for ‘Crepes in Paris’

SOURCE: Chamber of Progress ^[19]

FIGURE 2: French VPN Search Results for ‘Crepes in Paris’

SOURCE: Chamber of Progress ^[20]

Presumably, this is happening because a direct link to Google Maps would constitute “self-preferencing” wherein the search engine, Google, would be “unfairly” directing traffic to its own digital-navigation service. Such conduct is prohibited by Art. 6(5) of the DMA. But this kind of integration is very convenient for consumers, who can search for a restaurant and then quickly find the directions to walk or commute to it (and sometimes even book a table).

While removing some features, Google is also adding more results to its results pages, because it assumes that it is required under the DMA to provide “fair” links to competing sites like Yelp and TripAdvisor.^[21] In theory, the consequence of such requirements is “more options” for consumers. In practice, what consumers have is less relevant results, a more cluttered results page, and thus, a downgraded user experience.

Apple highlights another quality-degrading consequence of the DMA: the obligation imposed on platforms like iOS to allow competing app stores and to allow apps to be downloaded directly from their websites (commonly known as “sideloading”).^[22] This “openness,” however, would allow third-party applications to bypass controls and protections implemented to safeguard users’ security and privacy.^[23] This is already happening in Europe, where Apple has been forced to allow Epic Games to launch an alternative app store on its iOS operating system.[24] While this may seem a positive development for (some) developers and consumers, it could also harm user trust in the platform and thus decrease the total number of transactions, to the detriment of all parties involved (business users, consumers, and the owner of the platform). Indeed, “[p]hishers are using a novel technique to trick iOS and Android users into installing malicious apps that bypass safety guardrails built by both Apple and Google to prevent unauthorized apps.”[25] This sort of attack will be more effective in the absence of the protections in Apple’s App Store.[26]

The recent Microsoft/CrowdStrike outage—which affected many services around the world, including airlines and public services—is a good example of the tradeoffs between “openness” and the security and reliability of digital platforms. At least in part, the outage is explained by the fact that Microsoft must give access to its operating system to CrowdStrike and other developers. As explained in a Financial Times note:

Giving software companies that kind of access to an operating system is dangerous — it means you can quickly lose control of your computer if any of the software providers you rely on makes a mistake or is compromised. That is why Apple began informing third-party developers in 2020 that it would no longer grant them kernel-level access to the MacOS operating system (and also quite possibly why the CrowdStrike problem didn’t affect Apple devices).

But not all the fault lies with Microsoft. A 2009 agreement between the company and the European Commission requires it to grant outside developers the same access to Windows that its own security software has. The idea was to make it possible for other software companies to compete with Microsoft by ensuring many of its products and services are interoperable with outside software and tools. That’s a worthy goal, and many provisions in the agreement are entirely reasonable, such as requiring that Outlook support common calendar event and scheduling formats.

But the 2009 agreement is profoundly flawed in requiring Microsoft to make all of the APIs, or programming functions, that its own security software products use available to manufacturers of third-party security software products. This is the provision that requires Microsoft to give kernel-level access to companies such as CrowdStrike. Until it is changed, it’s not clear that Microsoft can implement the chief lesson of this debacle and start phasing out access, as Apple did four years ago.[27]

The problem with ex-ante regulations like the DMA, which prohibit some types of conduct absolutely, is that they don’t capture the complexities and tradeoffs generally present in market dynamics and, of course, specifically in digital markets. As Dirk Auer has pointed out, the alleged benefits of the DMA are looking more and more like broken promises:

When it was passed, European policymakers like Margrethe Vestager and Thierry Breton assured the public that the far-reaching regulation would not compromise security, lead to costlier services, or otherwise degrade users’ online experience. They also argued that it would be fast and easy to apply, thus avoiding the lengthy litigation that has come to be associated with competition enforcement.

As the effects of the DMA start to play out, however, these promises appear increasingly fanciful.

The biggest concern is that Europeans’ online safety is being compromised. Apple has warned that it will not be able to guarantee the safety of rival app stores and payment systems that can now access its ecosystem. If this sounds abstract, it is worth noting that these sorts of security flaws facilitated the Oct. 7 attacks carried out by Hamas. They also increase more mundane risks of identity theft and fraud.

Similarly, Amazon will struggle to exclude nefarious goods, sellers, and shippers from its online marketplace. Commenting on similar issues in the United States, the company surmised that it risked losing “customer trust by advertising something that is not a good deal for them.” This loss of consumer trust would, in turn, harm the bottom lines of the roughly two million businesses that rely on the platform.[28]

As indicated above, the DMA’s unintended consequences affect not only consumers, but also business users. Since Google began to implement the DMA in January 2024, early estimates suggest that clicks from Google ads to hotel websites fell by 17.6%.^[29] Presumably, this ought to be considered a failure, even by the DMA’s own (uncertain) standards.

Australia itself has experience with another regulation intended to address a “significant bargaining power imbalance” between digital platforms and businesses interacting with them. The News Media and Digital Platforms Mandatory Bargaining Code was enacted in 2021 to allow news organizations to collectively bargain with platforms like Google and Facebook for remuneration for news content featured on those platforms without breaching Australian competition laws.[30] It has, however, apparently resulted in less traffic for smaller media outlets and independent publishers. As reported by The Guardian: “Smaller publishers are already feeling the effects of a potential ban on news on Facebook, a parliamentary committee has heard, as news outlets small and large make the case for social media companies to be compelled to pay for news.”[31]

This is why, even when their relative lenity is considered, antitrust laws are more flexible and more likely to be appropriate for various markets and business models.^[32] As Posner illustrates, while rules are generally simpler and cheaper to enforce, they are often underinclusive or overinclusive. When facts like the mere size or number of users—which are disconnected from competitive goals—are determinative of their application, they are “especially apt to fail.”^[33]

All in all, as the ACCC acknowledges, most digital competition regulation “has only recently been enacted, and the full impact of recent international regulatory developments cannot yet be observed.”[34] In that vein, it would be wise to study markets, perform proper regulatory-impact analysis, and keep learning from the experience of other regulators and markets. It may be advisable, for instance, to follow the example of South Korea, which has hit the pause button on its proposal to regulate digital markets.[35]

Regarding major developments in digital-platform markets, the issues paper provides updated information about two markets: online private messaging and app marketplaces.

With respect to the online private-messaging market, the issues paper reminds that:

The ACCC’s September 2020 report found that both Facebook’s and Apple’s services benefited from identity-based network effects, providing them with significant competitive advantages over smaller suppliers of standalone services in Australia. However, because the use of Apple’s online private messaging services was limited to users of Apple devices, the ACCC found that iMessage was likely to impose weaker competitive constraints on Facebook Messenger and WhatsApp than those services imposed on iMessage.[36]

The paper also highlights that “Australians’ usage of these services is increasing. Based on recent ACMA survey data, in the 6 months prior to June 2023, 72% of Australian adults used Facebook Messenger (up from 66% in 2020), 51% used WhatsApp (up from 39% in 2020) and 35% used FaceTime (up from 33% in 2020).”[37]

It is important, at this point, to consider that, while the aforementioned messaging platforms (Facebook’s Messenger and Apple’s iMessage) have a large number of users and, therefore, a relevant market share and the advantage of network effects, this does not mean that the messaging-apps market is not competitive. The issues paper itself recognizes that:

Online private messaging services encompass a range of services, including text, audio and video messaging services, and are offered by a wide variety of platforms. The ACCC observed a wide range of online private messaging services available to Australian users, which were often highly differentiated, offering different features and functionalities, and used by consumers for a number of different purposes.[38]

Besides Facebook’s Messenger and Apple’s iMessage, several actors with smaller but not irrelevant market shares create competitive pressure on both.[39] Moreover, this is a market where consumers can use more than one application simultaneously (i.e., there is “multi-homing”). This allows consumers to try different services at little cost, which means that both current competitors and new entrants can challenge the market leaders. The fact that the use of these services is increasing in Australia provides an opportunity for these challengers.

Furthermore, although network effects offer an important advantage, they do not guarantee a comfortable monopolist position.[40] Catherine Tucker explains how network effects can be a double-edged sword, as they can also lead to a sudden decline of users to a marginally superior competitor.[41] Actual economic studies of data-network effects have been few and far between, with scant empirical evidence to support the theory that the control of personal data creates an insurmountable barrier to entry.^[42]

Regarding app marketplaces, the issues paper finds that:

In its March 2021 report, the ACCC found that Apple and Google were the predominant mobile app marketplace operators in Australia, offering the App Store and Play Store respectively.

The report referenced data from Statcounter, estimating that in December 2020, Apple iOS held 54% of the market share of mobile operating systems (OS) in Australia, while Android held 46%. Based on updated data from Statcounter, these shares appear to have remained stable – as of June 2024, Apple iOS again held 54% market share, while Android held 45%. The next closest competitor, Samsung, held less than 1% (0.86%).

In the March 2021 report, the ACCC considered it likely that Apple and Google held significant market power in mobile app distribution in Australia, due to their control of the iOS and Android mobile OS. This meant the App Store and the Play Store were ‘must haves’ for the majority of mobile app developers in Australia.[43]

The fact that both Apple and Google (Android) both have relatively high (and similar) market shares does not necessarily mean that the market is not competitive. Two powerful actors can discipline each other’s efforts to exercise market power.[44] Our point here is not that there cannot be any competition problem or anticompetitive conduct in this market, of course, but rather that more evidence than merely concentration or market shares is required in order to determine that there are even possible competition issues in this market. The rivalry between Apple and Android in the market has largely benefited consumers with innovation and increased access to a large array of features.[45]

III. Potential Emerging Issues: Competition and AI

Although the issues paper also identifies emerging issues in the online-gaming and cloud-computing markets, it identifies very specific (if somewhat theoretical) risks in the generative-AI market that echo the concerns raised by several competition agencies around the world. Drawing on a discussion paper by the Digital Platforms Regulators Forum, the issues paper notes that:

• Developing and operating large language models (LLMs) requires a large upfront financial investment, access to vast datasets, long development lead times, access to sophisticated AI systems and talent, and substantial ongoing computing costs and access to computing resources (such as cloud storage), which together create high barriers to entry.
• LLMs are likely to have features common to digital-platform services that make such markets tend toward concentration, including a positive feedback loop involving the collection and use of user data, economies of scale, and access to large volumes of high-quality user data. Because of these characteristics, new entrants could find it difficult to compete with digital-platform services that use LLMs as part of new and existing services.
• Generative AI could lead users to interact more with particular digital platforms. Over time, this may make it more difficult for users to leave these platforms. LLMs could allow big digital platforms to strengthen and expand their market power by continuing to engage in the allegedly anti-competitive practices the ACCC has previously observed. These include anti-competitive self-preferencing, tying, and restrictions on data access.[46]

We address these concerns in the following subsections. Subsection A summarizes recent calls for competition intervention in generative-AI markets. Subsection BB argues that many of these calls are underpinned by fears of data-related incumbency advantages (often referred to as “data-network effects”), including in the context of mergers. Subsection CC explains why these effects are unlikely to play a meaningful role in generative AI markets. Finally, subsection D offers five key takeaways to help policymakers better weigh the tradeoffs inherent in competition enforcement interventions in generative AI markets.

A. Calls for Intervention in AI Markets Should Acknowledge Actual Market Developments

It was once (and frequently) said that Google’s “data monopoly” was unassailable: “If ‘big data’ is the oil of the information economy, Google has Standard Oil-like monopoly dominance—and uses that control to maintain its dominant position.”[47] Similar claims of data dominance have been attached to nearly all large online platforms, including Facebook (Meta), Amazon, and Uber.[48]

While some of these claims continue even today (for example, “big data” is a key component of the U.S. Justice Department’s (DOJ) Google Search and adtech antitrust suits),[49] a shiny new data target has emerged in the form of generative AI. The launch of ChatGPT in November 2022, as well as the advent of AI image-generation services like Midjourney and Dall-E, have dramatically expanded the public’s conception of what is—and what might be—possible to achieve with generative-AI technologies built on massive datasets.

While these services remain both in the early stages of mainstream adoption and in the throes of rapid, unpredictable technological evolution, they are nevertheless already on the radar of competition policymakers around the world. Several antitrust enforcers appear to believe that, by acting now, they can avoid the “mistakes” that purportedly were made during the formative years of Web 2.0.[50] These mistakes, critics assert, include failing to appreciate the centrality of data in online markets, as well as letting mergers go unchecked and allowing early movers to entrench their market positions.^[51] As Federal Trade Commission (FTC) Chair Lina Khan has put it: “we are still reeling from the concentration that resulted from Web 2.0, and we don’t want to repeat the missteps of the past with AI.”^[52]

This response from the competition-policy world is deeply troubling. Rather than engage in critical self-assessment and adopt an appropriately restrained stance, the enforcement community appears to be champing at the bit. Rather than assessing their prior assumptions based on the current technological moment, enforcers’ top priority appears to be figuring out how to rapidly and almost reflexively deploy existing competition tools to address the presumed competitive failures presented by generative AI.^[53]

It has for some time been increasingly common for competition enforcers to argue that “data-network effects” serve not only to entrench incumbents in those markets where the data is collected, but also to confer similar, self-reinforcing benefits in adjacent markets. Several enforcers have, for example, prevented large online platforms from acquiring smaller firms in adjacent markets, citing the risk that they could use their vast access to data to extend their dominance into these new markets.^[54]

They have concomitantly launched consultations to ascertain the role that data plays in AI competition. For instance, in a recent consultation, the European Commission asked: “What is the role of data and what are its relevant characteristics for the provision of generative AI systems and/or components, including AI models?”^[55] The FTC has likewise been hypervigilant about the risks ostensibly posed by incumbents’ access to data. In comments submitted to the U.S. Copyright Office, for example, the FTC argued that:

The rapid development and deployment of AI also poses potential risks to competition. The rising importance of AI to the economy may further lock in the market dominance of large incumbent technology firms. These powerful, vertically integrated incumbents control many of the inputs necessary for the effective development and deployment of AI tools, including cloud-based or local computing power and access to large stores of training data. These dominant technology companies may have the incentive to use their control over these inputs to unlawfully entrench their market positions in AI and related markets, including digital content markets.[56]

Likewise, Jonathan Kanter, assistant U.S. attorney general for antitrust, recently claimed that:

We also see structures and trends in AI that should give us pause AI relies on massive amounts of data and computing power, which can give already dominant firms a substantial advantage. Powerful networks and feedback effects may enable dominant firms to control these new markets, and existing power in the digital economy may create a powerful incentive to control emerging innovations that will not only impact our economy, but the health and well-being of our society and free expression itself.[57]

On an even more hyperbolic note, Andreas Mundt, the head of Germany’s Federal Cartel Office, called AI a “first-class fire accelerator” for anticompetitive behavior and argued it “will make all the problems only worse.”[58] He further argued that “there’s a great danger that we’ll get an even deeper concentration of digital markets and power increase at various levels, from chips to the front end.”[59] In short, Mundt is one of many policymakers who believe that AI markets will enable incumbent tech firms to further entrench their market positions.

Certainly, it makes sense that the largest online platforms—including Alphabet, Meta, Apple, and Amazon—should have a meaningful advantage in the burgeoning markets for generative AI services. After all, it is widely recognized that data is an essential input for generative AI.^[60] This competitive advantage should be all the more significant, given that these firms have been at the forefront of AI technology for more than a decade. Over this period, Google’s DeepMind and AlphaGo and Meta’s NLLB-200 have routinely made headlines.[61] Apple and Amazon also have vast experience with AI assistants, and all of these firms deploy AI technologies throughout their platforms.[62]

Contrary to what one might expect, however, the tech giants have, to date, been largely unable to leverage their vast troves of data to outcompete startups like OpenAI and Midjourney. At the time of writing, for instance, OpenAI’s ChatGPT appears to be, by far, the most successful chatbot,[63] despite the large tech platforms’ apparent access to far more (and more up-to-date) data.

There are important lessons to glean from these developments, if only enforcers would stop to reflect. The meteoric rise of consumer-facing AI services should offer competition enforcers and policymakers an opportunity for introspection. As we explain, the rapid emergence of generative AI may undercut many core assumptions of today’s competition-policy debates, which have focused largely on the rueful after-effects of the purported failure of 20^th-century antitrust to address the alleged harms of 21^st-century technology. These include the notions that data advantages constitute barriers to entry and can be leveraged to project dominance into adjacent markets; that scale itself is a market failure to be addressed by enforcers; and that the use of consumer data is inherently harmful to those consumers.

B. Data-Network Effects Theory and Enforcement

Proponents of more extensive intervention by competition enforcers into digital markets often cite data-network effects as a source of competitive advantage and a barrier to entry (although terms like “economies of scale and scope” may offer more precision).^[64] The crux of the argument is that “the collection and use of data creates a feedback loop of more data, which ultimately insulates incumbent platforms from entrants who, but for their data disadvantage, might offer a better product.”^[65] This self-reinforcing cycle purportedly leads to market domination by a single firm. Thus, it is argued, e.g., that Google’s “ever-expanding control of user personal data, and that data’s critical value to online advertisers, creates an insurmountable barrier to entry for new competition.”^[66]

But it is important to note the conceptual problems these claims face. Because data can be used to improve products’ quality and/or to subsidize their use, if possessing data constitutes an entry barrier, then any product improvement or price reduction made by an incumbent could be problematic. This is tantamount to an argument that competition itself is a cognizable barrier to entry. Of course, it would be a curious approach to antitrust if competition were treated as a problem, as it would imply that firms should under-compete—i.e., should forego consumer-welfare enhancements—in order to inculcate a greater number of firms in a given market, simply for its own sake.[67]

Meanwhile, actual economic studies of data-network effects have been few and far between, with scant empirical evidence to support the theory.^[68] Andrei Hagiu and Julian Wright’s theoretical paper offers perhaps the most comprehensive treatment of the topic to date.[69] The authors ultimately conclude that data-network effects can be of differing magnitudes and have varying effects on firms’ incumbency advantage.^[70] They cite Grammarly (an AI writing-assistance tool) as a potential example: “As users make corrections to the suggestions offered by Grammarly, its language experts and artificial intelligence can use this feedback to continue to improve its future recommendations for all users.”^[71]

This is echoed by economists who contend that “[t]he algorithmic analysis of user data and information might increase incumbency advantages, creating lock-in effects among users and making them more reluctant to join an entrant platform.”^[72] Crucially, some scholars take this logic a step further, arguing that platforms may use data from their “origin markets” in order to enter and dominate adjacent ones:

First, as we already mentioned, data collected in the origin market can be used, once the enveloper has entered the target market, to provide products more efficiently in the target market. Second, data collected in the origin market can be used to reduce the asymmetric information to which an entrant is typically subject when deciding to invest (for example, in R&D) to enter a new market. For instance, a search engine could be able to predict new trends from consumer searches and therefore face less uncertainty in product design.^[73]

This possibility is also implicit in Hagiu and Wright’s paper.[74] Indeed, the authors’ theoretical model rests on an important distinction between “within-user” data advantages (that is, having access to more data about a given user) and “across-user” data advantages (information gleaned from having access to a wider user base). In both cases, there is an implicit assumption that platforms may use data from one service to gain an advantage in another market (because what matters is information about aggregate or individual user preferences, regardless of its origin).

Our review of the economic evidence suggests that several scholars have, with varying degrees of certainty, raised the possibility that incumbents may leverage data advantages to stifle competitors in their primary market or in adjacent ones (be it via merger or organic growth). As we explain below, however, there is ultimately little evidence to support such claims. Policymakers have nonetheless been keenly receptive to these limited theoretical findings, basing multiple decisions on these theories, often with little consideration given to the caveats that accompany them.^[75]

Indeed, it is remarkable that, in its section on “[t]he data advantage for incumbents,” the “Furman Report” created for the UK government cited only two empirical economic studies, and they offer directly contradictory conclusions with respect to the strength of data advantages.[76] The report nevertheless concluded that data “may confer a form of unmatchable advantage on the incumbent business, making successful rivalry less likely,”[77] and it adopted without reservation what it deemed “convincing” evidence from non-economists that have no apparent empirical basis.[78]

In the Google/Fitbit merger proceedings, the European Commission found that the combination of data from Google services with that of Fitbit devices would reduce competition in advertising markets:

Giving [sic] the large amount of data already used for advertising purposes that Google holds, the increase in Google’s data collection capabilities, which goes beyond the mere number of active users for which Fitbit has been collecting data so far, the Transaction is likely to have a negative impact on the development of an unfettered competition in the markets for online advertising.^[79]

As a result, the Commission cleared the merger only on the condition that Google refrain from using data from Fitbit devices for its advertising platform.^[80] The Commission also appears likely to focus on similar issues in its ongoing investigation of Microsoft’s investment in OpenAI.[81]

Along similar lines, in its complaint to enjoin Meta’s purchase of Within Unlimited—makers of the virtual-reality (VR) fitness app Supernatural—the FTC relied on, among other things, the fact that Meta could leverage its data about VR-user behavior to inform its decisions and potentially outcompete rival VR-fitness apps: “Meta’s control over the Quest platform also gives it unique access to VR user data, which it uses to inform strategic decisions.”^[82]

The DOJ’s twin cases against Google also implicate data leveraging and data barriers to entry. The agency’s adtech complaint charges that “Google intentionally exploited its massive trove of user data to further entrench its monopoly across the digital advertising industry.”[83] Similarly, in its Google Search complaint, the agency argued that:

Google’s anticompetitive practices are especially pernicious because they deny rivals scale to compete effectively. General search services, search advertising, and general search text advertising require complex algorithms that are constantly learning which organic results and ads best respond to user queries; the volume, variety, and velocity of data accelerates the automated learning of search and search advertising algorithms.[84]

Finally, updated merger guidelines published in recent years by several competition enforcers cite the acquisition of data as a potential source of competition concerns. For instance, the FTC and DOJ’s 2023 guidelines state that “acquiring data that helps facilitate matching, sorting, or prediction services may enable the platform to weaken rival platforms by denying them that data.”^[85] Likewise, the UK Competition and Markets Authority warned against incumbents acquiring firms in order to obtain their data and foreclose other rivals:

Incentive to foreclose rivals…

7.19(e) Particularly in complex and dynamic markets, firms may not focus on short term margins but may pursue other objectives to maximise their long-run profitability, which the CMA may consider. This may include… obtaining access to customer data….^[86]

In short, competition authorities around the globe have taken an increasingly aggressive stance on data-network effects. Among the ways this has manifested is in enforcement decisions based on fears that data a platform collects in one market might confer decisive competitive advantages in adjacent markets. Unfortunately, these concerns rest on little to no empirical evidence, either in the economic literature or the underlying case records.

C. Data-Incumbency Advantages in Generative AI

Given the assertions detailed in the previous section, it would be reasonable to assume that firms such as Google, Meta, and Amazon should be in pole position to meet the burgeoning demand for generative AI. After all, these firms have not only been at the forefront of the field for the better part of a decade, but they also have access to vast troves of data, the likes of which their rivals could only dream when they launched their own services. Thus, the authors of the Furman Report caution that “to the degree that the next technological revolution centres around artificial intelligence and machine learning, then the companies most able to take advantage of it may well be the existing large companies because of the importance of data for the successful use of these tools.”[87]

To date, however, this is not how things have unfolded (although it bears noting that these technologies remain in flux and the competitive landscape is susceptible to change). The first significantly successful generative AI service was arguably not from either Meta—which had been working on chatbots for years and had access to, arguably, the world’s largest database of actual chats—or Google. Instead, the breakthrough came from a previously unknown firm called OpenAI.

OpenAI’s ChatGPT service currently accounts for an estimated 60% of visits to online AI tools (although reliable numbers are somewhat elusive).[88] It broke the record for the fastest online service to reach 100 million users (in only a couple of months), more than four times faster than TikTok, the previous record holder.^[89] Based on Google Trends data, ChatGPT is nine times more popular worldwide than Google’s own Bard service, and 14 times more popular in the United States.^[90] In April 2023, ChatGPT reportedly registered 206.7 million unique visitors, compared to 19.5 million for Google’s Bard.^[91] In short, at the time we are writing, ChatGPT appears to be the most popular chatbot. The entry of large players such as Google Bard or Meta AI appear to have had little effect thus far on its leading position.^[92]

The picture is similar in the field of AI image generation. As of August 2023, Midjourney, Dall-E, and Stable Diffusion appear to be the three market leaders in terms of user visits.^[93] This is despite competition from the likes of Google and Meta, who arguably have access to unparalleled image and video databases by virtue of their primary platform activities.^[94]

This raises several crucial questions: how have these AI upstarts managed to be so successful, and is their success just a flash in the pan before Web 2.0 giants catch up and overthrow them? While we cannot answer either of these questions dispositively, we offer what we believe to be some relevant observations concerning the role and value of data in digital markets.

A first important observation is that empirical studies suggest that data exhibits diminishing marginal returns. In other words, past a certain point, acquiring more data does not confer a meaningful edge to the acquiring firm. As Catherine Tucker put it, following a review of the literature: “Empirically there is little evidence of economies of scale and scope in digital data in the instances where one would expect to find them.”^[95]

Likewise, following a survey of the empirical literature on this topic, Geoffrey Manne and Dirk Auer conclude that:

Available evidence suggests that claims of “extreme” returns to scale in the tech sector are greatly overblown. Not only are the largest expenditures of digital platforms unlikely to become proportionally less important as output increases, but empirical research strongly suggests that even data does not give rise to increasing returns to scale, despite routinely being cited as the source of this effect.^[96]

In other words, being the firm with the most data appears to be far less important than having enough data. Moreover, this lower bar may be accessible to far more firms than one might initially think possible. Furthermore, obtaining sufficient data could become easier still—that is, the volume of required data could become even smaller—with technological progress. For instance, synthetic data may provide an adequate substitute to real-world data,[97] or may even outperform real-world data.[98] As Thibault Schrepel and Alex Pentland surmise:

[A]dvances in computer science and analytics are making the amount of data less relevant every day. In recent months, important technological advances have allowed companies with small data sets to compete with larger ones.^[99]

Indeed, past a certain threshold, acquiring more data might not meaningfully improve a service, where other improvements (such as better training methods or data curation) could have a large impact. In fact, there is some evidence that excessive data impedes a service’s ability to generate results appropriate for a given query: “[S]uperior model performance can often be achieved with smaller, high-quality datasets than massive, uncurated ones. Data curation ensures that training datasets are devoid of noise, irrelevant instances, and duplications, thus maximizing the efficiency of every training iteration.”[100]

Consider, for instance, a user who wants to generate an image of a basketball. Using a model trained on an indiscriminate range and number of public photos in which a basketball appears surrounded by copious other image data, the user may end up with an inordinately noisy result. By contrast, a model trained with a better method on fewer, more carefully selected images could readily yield far superior results.[101] In one important example:

The model’s performance is particularly remarkable, given its small size. “This is not a large language model trained on the whole Internet; this is a relatively small transformer trained for these tasks,” says Armando Solar-Lezama, a computer scientist at the Massachusetts Institute of Technology, who was not involved in the new study…. The finding implies that instead of just shoving ever more training data into machine-learning models, a complementary strategy might be to offer AI algorithms the equivalent of a focused linguistics or algebra class.[102]

Platforms’ current efforts are thus focused on improving LLMs’ mathematical and logical reasoning, rather than maximizing the size of training datasets.^[103] Two points stand out. The first is that firms like OpenAI rely largely on publicly available datasets—such as GSM8K—to train their LLMs.^[104] Second, the real challenge to creating innovative AI lies not so much in collecting data, but in creating innovative AI-training processes and architectures:

[B]uilding a truly general reasoning engine will require a more fundamental architectural innovation. What’s needed is a way for language models to learn new abstractions that go beyond their training data and have these evolving abstractions influence the model’s choices as it explores the space of possible solutions.

We know this is possible because the human brain does it. But it might be a while before OpenAI, DeepMind, or anyone else figures out how to do it in silicon.^[105]

Furthermore, it is worth noting that the data most relevant to startups in a given market may not be those held by large incumbent platforms in other markets. They might instead be data specific to the market in which the startup is active or, even better, to the given problem it is attempting to solve:

As Andres Lerner has argued, if you wanted to start a travel business, the data from Kayak or Priceline would be far more relevant. Or if you wanted to start a ride-sharing business, data from cab companies would be more useful than the broad, market-cross-cutting profiles Google and Facebook have. Consider companies like Uber, Lyft and Sidecar that had no customer data when they began to challenge established cab companies that did possess such data. If data were really so significant, they could never have competed successfully. But Uber, Lyft and Sidecar have been able to effectively compete because they built products that users wanted to use—they came up with an idea for a better mousetrap. The data they have accrued came after they innovated, entered the market and mounted their successful challenges—not before.[106]

The bottom line is that data is not the be-all and end-all that many in competition circles make it out to be. While data may often confer marginal benefits, there is little evidence that these benefits are ultimately decisive.[107] As a result, incumbent platforms’ access to vast numbers of users and troves of data in their primary markets might only marginally affect their competitiveness in AI markets.

A related observation is that firms’ capabilities and other features of their products arguably play a more important role than the data they own.[108] Examples of this abound in digital markets. Google overthrew Yahoo in search, despite initially having access to far fewer users and far less data. Google and Apple overcame Microsoft in the smartphone operating-system market, despite having comparatively tiny ecosystems (at the time) to leverage. TikTok rose to prominence despite intense competition from incumbents like Instagram, which had much larger userbases. In each of these cases, important product-design decisions (such as the PageRank algorithm, recognizing the specific needs of mobile users,^[109] and TikTok’s clever algorithm) appear to have played far more significant roles than the firms’ initial user and data endowments (or lack thereof).

All of this suggests that the early success of OpenAI likely has more to do with its engineering decisions than with what data it did or did not possess. Going forward, OpenAI and its rivals’ relative abilities to offer and monetize compelling use cases by offering custom versions of their generative AI technologies will arguably play a much larger role than (and contribute to) their ownership of data.[110] In other words, the ultimate challenge is arguably to create a valuable platform, of which data ownership is a consequence, not a cause.

It is also important to note that, in those instances where it is valuable, data does not just fall from the sky. Instead, it is through smart business and engineering decisions that firms can generate valuable information (which does not necessarily correlate with owning more data). For instance, OpenAI’s success with ChatGPT is often attributed to its more efficient algorithms and training models, which arguably have enabled the service to improve more rapidly than its rivals.^[111] Likewise, the ability of firms like Meta and Google to generate valuable data for advertising arguably depends more on design decisions that elicit the right data from users, rather than the raw number of users in their networks.

Put differently, setting up a business so as to gather and organize the right information is more important than simply owning vast troves of data.[112] Even in those instances where high-quality data is an essential parameter of competition, it does not follow that having vaster databases or more users on a platform necessarily leads to better information for the platform. Indeed, if data ownership consistently conferred a significant competitive advantage, these new AI firms would not be where they are today.

Moreover, it is important not to neglect the role that open-source models currently play in fostering innovation and competition. As former DOJ Chief Antitrust Economist Susan Athey pointed out in a recent interview, “[the AI industry] may be very concentrated, but if you have two or three high quality—and we have to find out what that means, but high enough quality—open models, then that could be enough to constrain the for-profit LLMs.”[113] Open-source models are important because they allow innovative startups to build upon models already trained on large datasets—therefore entering the market without incurring that initial cost. Apparently, there is no lack of open-source models, since companies like xAI, Meta, and Google offer their AI models for free.[114]

This does not, of course, mean that data is worthless. Rather, it means that competition authorities should not assume that the mere possession of data is a dispositive competitive advantage, absent compelling empirical evidence to support such a finding. In this light, the current wave of decisions and competition-policy pronouncements that rely on data-related theories of harm are premature.

D. Five Key Takeaways: Reconceptualizing the Role of Data in Generative-AI Competition

As we explain above, data-network effects are not the source of barriers to entry that they are sometimes made out to be. The picture is far more nuanced. Indeed, as economist Andres Lerner demonstrated almost a decade ago (and the assessment is only truer today):

Although the collection of user data is generally valuable for online providers, the conclusion that such benefits of user data lead to significant returns to scale and to the entrenchment of dominant online platforms is based on unsupported assumptions. Although, in theory, control of an “essential” input can lead to the exclusion of rivals, a careful analysis of real-world evidence indicates that such concerns are unwarranted for many online businesses that have been the focus of the “big data” debate.[115]

While data can be an important part of the competitive landscape, incumbents’ data advantages are far less pronounced than today’s policymakers commonly assume. In that respect, five primary lessons emerge:

1. Data can be (very) valuable, but beyond a certain threshold, those benefits tend to diminish. In other words, having the most data is less important than having enough;
2. The ability to generate valuable information does not depend on the number of users or the amount of data a platform has previously acquired;
3. The most important datasets are not always proprietary;
4. Technological advances and platforms’ engineering decisions affect their ability to generate valuable information, and this effect swamps those that stem from the amount of data they own; and
5. How platforms use data is arguably more important than what data or how much data they own.

These lessons have important ramifications for policy debates over the competitive implications of data in technologically evolving areas.

First, it is not surprising that startups, rather than incumbents, have taken an early lead in generative AI (and in Web 2.0 before it). After all, if data-incumbency advantages are small or even nonexistent, then smaller and more nimble players may have an edge over established tech platforms. This is all the more likely given that, despite significant efforts, the biggest tech platforms were unable to offer compelling generative-AI chatbots and image-generation services before the emergence of ChatGPT, Dall-E, Midjourney, etc.

This suggests that, in a process akin to Clayton Christensen’s “innovator’s dilemma,”[116] something about the incumbent platforms’ existing services and capabilities might have been holding them back in this emerging industry. Of course, this does not necessarily mean that those same services or capabilities could not become an advantage when the generative-AI industry starts addressing issues of monetization and scale.[117] But it does mean that assumptions about a firm’s market power based primarily on its possession of data are likely to be off the mark.

Another important implication is that, paradoxically, policymakers’ efforts to prevent Web 2.0 platforms from competing freely in generative-AI markets may ultimately backfire and lead to less, not more, competition. Indeed, OpenAI is currently acquiring a sizeable lead in generative AI. While competition authorities might like to think that other startups will emerge and thrive in this space, it is important not to confuse those desires with reality. While there currently exists a vibrant AI-startup ecosystem, there is at least a case to be made that significant competition for today’s AI leaders will come from incumbent Web 2.0 platforms—although nothing is certain at this stage.

Policymakers should beware not to stifle that competition on the misguided assumption that competitive pressure from large incumbents is somehow less valuable to consumers than those that originate from smaller firms. This is particularly relevant in the context of merger control. An acquisition (or an “acqui-hire”) by a “Big Tech” company does not only, in principle, entail a minor risk to harm competition (it is not a horizontal merger),[118] but could create a stronger competitor to the current market leaders.

Finally, even if there were a competition-related market failure to be addressed in the field of generative AI (which is anything but clear), the remedies under contemplation may do more harm than good. Some of the solutions that have been put forward have highly ambiguous effects on consumer welfare. Scholars have shown that, e.g., mandated data sharing—a solution championed by EU policymakers, among others—may sometimes dampen competition in generative AI.[119] This is also true of legislation like the General Data Protection Regulation (GDPR), which makes it harder for firms to acquire more data about consumers—assuming such data is, indeed, useful to generative AI services.[120]

In sum, it is a flawed understanding of the economics and practical consequences of large agglomerations of data that has led competition authorities to believe data-incumbency advantages are likely to harm competition in generative AI—or even in the data-intensive Web 2.0 markets that preceded it. Indeed, competition or regulatory intervention to “correct” data barriers and data network and scale effects is liable to do more harm than good.

[1] Australian Competition & Consumer Commission, Digital Platform Services Inquiry —March 2025 — Final Report Issues Paper (25 Jul. 25 2024), available at https://www.accc.gov.au/system/files/dpsi-10-final-report-issues-paper.pdf?ref=0&download=y (hereinafter “issues paper” or “final report issues paper”). See also Press Release, Final Digital Platforms Report to Focus on Global Developments and Emerging Competition and Consumer Issues, Australian Competition & Consumer Commission (25 Jul. 2024), https://www.accc.gov.au/media-release/final-digital-platforms-report-to-focus-on-global-developments-and-emerging-competition-and-consumer-issues.

[2] Press release, id.

[3] Digital Platform Services Inquiry: Interim Report No. 5 — Regulatory Reform, Australian Competition & Consumer Commission (Sep. 2022), at 47, available at https://www.accc.gov.au/system/files/Digital%20platform%20services%20inquiry%20-%20September%202022%20interim%20report.pdf.

[4] Artificial intelligence is, of course, not a market (at least, not an antitrust-relevant market). Within the realm of what is called “AI,” companies offer myriad products and services, and specific relevant markets would need to be defined before assessing harm to competition in specific cases.

[5] Issues paper, supra note 1, at 4.

[6] Robert Armstrong & Ethan Wu, What Big Tech Antitrust Gets Wrong. An Interview with Herbert Hovenkamp, Financial Times (19 Jan. 2024), https://www.ft.com/content/4eec8bc3-c892-4704-ae66-a4432c6d4fd7.

[7] By ex-ante regulation, we mean specific rules and duties that are sector-specific (“digital markets”), whose application would not be based on the effects of the conduct regulated, and where fines would apply in case of non-compliance. See Bruce H. Kobayashi & Joshua D. Wright, Antitrust and Ex-Ante Sector Regulation, The GAI Report on the Digital Economy 25 (2020); see also id. at Table 1 at 869.

[8] See Robert Cooter & Tomas Ulen, Law and Economics 40-43 (2000); W. Kip Viscusi, Joseph E. Harrington, Jr. and John M. Vernon, Economics of Regulation and Antitrust 376-79 (2005).

[9] David S. Evans & Richard Schmalensee, Debunking The “Network Effects” Bogeyman, Regulation 36, 39 (Winter 2017-2018) available at https://www.cato.org/sites/cato.org/files/serials/files/regulation/2017/12/regulation-v40n4-1.pdf.

[10] Giuseppe Colangelo, Evaluating the Case for Regulation of Digital Platforms, The GAI Report on the Digital Economy 905, 930 (2020) https://gaidigitalreport.com/2020/10/04/evaluating-the-case-for-ex-ante-regulation-of-digital-platforms.

[11] Interim Report No 5, supra note 3, at 48.

[12] Australian Competition & Consumer Commission, supra note 3, at 48-49.

[13] We often run the risk of condemning business practices and models we don’t fully understand. Sometimes, even the businesses that implement them don’t fully know or understand the impact of such practices. See Frank H. Easterbrook, The Limits of Antitrust, 63 Tex. L. Rev. 1 (1984).

[14] Dirk Auer & Lazar Radic, The Growing Legacy of Intel, 14 J. Eur. Comp. L. & Pract. 15 (2023), (“Competition cases routinely hinge on the fundamental distinction between conduct that anti-competitively serves to exclude competitors, on the one hand, and competition on the merits that may lead firms to exit the market, on the other. (…) anticompetitive foreclosure and competition on the merits both ultimately result in the same observable outcome: namely, that rivals exit the market. In order to draw the line, policymakers must infer both the root causes and the effects of firms’ market exit”.)

[15] Jean Tirole, Competition and the Industrial Challenge for the Digital Age, 15 Annual Rev. of Econ. 573, 581 (2023), available at https://www.annualreviews.org/content/journals/10.1146/annurev-economics-090622-024222.

[16] Alba Ribera, La Regulación de los Ecosistemas Digitales Frente a las Relaciones Complejas se los Operadores Económicos, Centro Competencia (18 Oct. 2023), https://centrocompetencia.com/regulacion-ecosistemas-digitales-relaciones-complejas-operadores-economicos (Free translation of the following text in Spanish: “Uno de los mayores ejemplos de la dicotomía que se erige entre los distintos tipos de consecuencias que se pueden generar por la captura regulatoria de los ecosistemas digitales lo podemos encontrar en la reciente decisión de Meta, de no lanzar su nuevo servicio Threads en el Espacio Económico Europeo. En la medida en que su servicio podría interpretarse de forma que cayera dentro de la definición de un “servicio básico de plataforma” perteneciente a la categoría de redes sociales en línea” (listada por la LMD), Meta decidió abstenerse de entrar en el mercado europeo, por la carga desproporcionada que le supondría las exigentes obligaciones impuestas por la LMD. Cabe notar que Threads es aún un servicio entrante en el mercado de redes sociales en línea, en contraste con la posición predominante ocupada por la actual X (anteriormente conocida como Twitter). De esta forma, observamos que la categorización como servicio básico de plataforma unifica y elimina todos los matices que el propio juego de la libre competencia opera respecto de servicios entrantes en los mercados.”).

[17] Press Release, Facebook and Instagram to Offer Subscription for No Ads in Europe, Meta (30 Oct. 2023), https://about.fb.com/news/2023/10/facebook-and-instagram-to-offer-subscription-for-no-ads-in-europe.

[18] Geoffrey Manne & Dirk Auer, Brussels Effect or Brussels Defect: Digital Regulation in Emerging Markets, Truth on the Market (20 Dec. 2022), https://truthonthemarket.com/2022/12/20/brussels-effect-or-brussels-defect-digital-regulation-in-emerging-markets.

[19] Adam Kovacevich, Europe’s Digital Market Act Fails Consumers, Chamber of Progress (4 Mar. 2024), https://medium.com/chamber-of-progress/europes-digital-market-act-fails-consumers-dcaf70cc548c.

[20] Id.

[21] Id.

[22] See Jon Porter & David Pierce, Apple Is Bringing Sideloading and Alternate App Stores to the iPhone, The Verge (25 Jan. 2024), https://www.theverge.com/2024/1/25/24050200/apple-third-party-app-stores-allowed-iphone-ios-europe-digital-markets-act.

[23] See Apple, Complying with the Digital Markets Act (2024), available at https://developer.apple.com/security/complying-with-the-dma.pdf.

[24] Kim Mackrael, Apple’s Hold on the App Store Is Loosening, at Least in Europe, Wall St. J. (16 Aug. 2024), https://www.wsj.com/tech/epic-games-apple-app-store-europe-44ceda50.

[25] Dan Goodin, Novel Technique Allows Malicious Apps to Escape iOS and Android Guardrails, ArsTechnica (21 Aug. 2024), https://arstechnica.com/security/2024/08/novel-technique-allows-malicious-apps-to-escape-ios-and-android-guardrails.

[26] See id. (“Both mobile operating systems employ mechanisms designed to help users steer clear of apps that steal their personal information, passwords, or other sensitive data. iOS bars the installation of all apps other than those available in its App Store, an approach widely known as the Walled Garden.”).

[27] Josephine Wolff, Software crash Exposes Tensions Between Security and Competition, Financial Times (28 Jul. 2024), https://www.ft.com/content/60dde560-194a-40d1-8c98-1d96d6d019a0.

[28] Dirk Auer, The Broken Promises of Europe’s Digital Regulation, Truth on the Market (12 Mar. 2024), https://truthonthemarket.com/2024/03/12/the-broken-promises-of-europes-digital-regulation.

[29] Mirai, DMA’s Negative Impact (Feb. 2024), available at LinkedIn, https://www.linkedin.com/feed/update/urn:li:activity:7161330551709138945 (“Since the implementation of the DMA on the 19th of January, the number of clicks from Google Hotel Ads to hotel websites has decreased by 17.6% in EU countries compared to the rest of the world.”).

[30] See Press Release, Parliament Passes News Media and Digital Platforms Mandatory Bargaining Code, Treasury Portfolio Ministers and Minister for Communications, Urban Infrastructure, Cities and the Arts, (25 Feb. 2021), https://ministers.treasury.gov.au/ministers/josh-frydenberg-2018/media-releases/parliament-passes-news-media-and-digital-platforms.

[31] Josh Taylor, Facebook’s Potential News Ban Already Affecting Smaller Australian Media Outlets, Inquiry Told, The Guardian (21 Jun. 2024), https://www.theguardian.com/media/article/2024/jun/21/facebooks-potential-news-ban-already-affecting-smaller-australian-media-outlets-inquiry-told.

[32] See, e.g., Thomas Lambert, Tech Platforms and Market Power: What’s the Optimal Policy Response?, Mercatus Working Paper (Nov. 2021), at 14, available at https://www.mercatus.org/research/working-papers/tech-platforms-and-market-power-whats-optimal-policy-response (“Because they are more rigid and prescriptive than antitrust’s flexible standards, and thus less likely to be appropriate for a broad range of diverse firms, ex ante rules addressing market power concerns tend to be limited in scope. They are usually tailored for a particular industry or group of firms. Antitrust’s standards are focused on ends rather than specific means, and are therefore less likely to ‘misfire’ when applied broadly.”).

[33] See Richard Posner, Antitrust Law (2nd. ed. 2001), at 39 (“Rules are generally simpler and cheaper to enforce than standards and provide clearer guidance both to the people subject to them and to the courts that administer them. But they are often either underinclusive or overinclusive, and sometimes they are both at the same time. They are especially apt to fail as a sensible method of lawmaking when the relation of the rule’s purpose to the fact of facts that it makes determinative or legality is unclear. In such cases, the decision whether to characterize the case as falling within the domain of the rule may depend on the same factors that would determine legality under a standard.”)

[34] Issues paper, supra note 1, at 8.

[35] Kwon Soon-wan &Yeom Hyun-a, South Korea Hits Pause on Anti-Monopoly Platform Act Targeting Google, Apple, The Chosun Daily (8 Feb. 2024), https://www.chosun.com/english/national-en/2024/02/08/A4U4X6TWEFFOXF7ITCS5K6SZN4.

[36] Issues paper, supra note 1, at 10.

[37] Id. at 9-10.

[38] Id. at 9 (emphasis added).

[39] See, Most Used Messenger by Brand in Australia as of June 2024, Statista (31 Jul. 2024), https://www.statista.com/forecasts/1187986/most-used-messenger-by-brand-in-australia.

[40] See supra note 9 and accompanying text.

[41] Catherine Tucker, Network Effects and Market Power: What Have We Learned in the Last Decade?, Antitrust 72, 75-76 (2018) (“the rise and fall of such platforms may also be more dramatic and renders such platforms far more vulnerable to a marginally superior competitor. The sudden decline of MySpace can be explained by the idea that given limited ability to spend time on two similar social networks, users switched far more quickly to a competitor.”).

^[42] For a review of the literature on increasing returns to scale in data (a topic somewhat broader than data-network effects), see Geoffrey Manne & Dirk Auer, Antitrust Dystopia and Antitrust Nostalgia: Alarmist Theories of Harm in Digital Markets and Their Origins, 28 Geo. Mason L. Rev. 1281, 1344 (2021).

[43] Issues paper, supra note 1, at 10.

[44] See generally William J. Baumol, Contestable Markets: An Uprising in the Theory of Industry Structure, 72 Am. Econ. Rev. 1 (1982); William J. Baumol, John Panzar & Robert D. Willig, Contestable Markets and the Theory of Industry Structure (revised ed. 1988).

[45] See, e.g., Andrew Lanxon, Android vs. iPhone: 15 Years of Innovation Through Rivalry, CNET (24 Apr. 2024), https://www.cnet.com/tech/mobile/smartphone-showdown-15-years-of-android-vs-iphone.

[46] Issues paper, supra note 1, at 13-14.

[47] Nathan Newman, Taking on Google’s Monopoly Means Regulating Its Control of User Data, Huffington Post (24 Sep. 2013), http://www.huffingtonpost.com/nathan-newman/taking-on-googlesmonopol_b_3980799.html.

[48] See, e.g., Lina Khan & K. Sabeel Rahman, Restoring Competition in the U.S. Economy, in Untamed: How to Check Corporate, Financial, and Monopoly Power (Nell Abernathy, Mike Konczal, & Kathryn Milani, eds., 2016), at 23. (“From Amazon to Google to Uber, there is a new form of economic power on display, distinct from conventional monopolies and oligopolies…, leverag[ing] data, algorithms, and internet-based technologies… in ways that could operate invisibly and anticompetitively.”); Mark Weinstein, I Changed My Mind—Facebook Is a Monopoly, Wall St. J. (1 Oct. 2021), https://www.wsj.com/articles/facebook-is-monopoly-metaverse-users-advertising-platforms-competition-mewe-big-tech-11633104247 (“[T]he glue that holds it all together is Facebook’s monopoly over data…. Facebook’s data troves give it unrivaled knowledge about people, governments—and its competitors.”).

[49] See, generally, Abigail Slater, Why “Big Data” Is a Big Deal, The Reg. Rev. (6 Nov. 2023), https://www.theregreview.org/2023/11/06/slater-why-big-data-is-a-big-deal; Amended Complaint at ¶36, United States v. Google, 1:20-cv-03010- (D.D.C. 2020); Complaint at ¶37, United States v. Google, 1:23-cv-00108 (E.D. Va. 2023), https://www.justice.gov/opa/pr/justice-department-sues-google-monopolizing-digital-advertising-technologies (“Google intentionally exploited its massive trove of user data to further entrench its monopoly across the digital advertising industry.”).

[50] See, e.g., Press Release, Commission Launches Calls for Contributions on Competition in Virtual Worlds and Generative AI, European Commission (9 Jan. 2024), https://ec.europa.eu/commission/presscorner/detail/en/IP_24_85; Krysten Crawford, FTC’s Lina Khan Warns Big Tech over AI, SIEPR (3 Nov. 2020), https://siepr.stanford.edu/news/ftcs-lina-khan-warns-big-tech-over-ai (“Federal Trade Commission Chair Lina Khan delivered a sharp warning to the technology industry in a speech at Stanford on Thursday: Antitrust enforcers are watching what you do in the race to profit from artificial intelligence.”) (emphasis added).

[51] See, e.g., John M. Newman, Antitrust in Digital Markets, 72 Vand. L. Rev. 1497, 1501 (2019) (“[T]he status quo has frequently failed in this vital area, and it continues to do so with alarming regularity. The laissez-faire approach advocated for by scholars and adopted by courts and enforcers has allowed potentially massive harms to go unchecked.”); Bertin Martins, Are New EU Data Market Regulations Coherent and Efficient?, Bruegel Working Paper 21/23 (2023), https://www.bruegel.org/working-paper/are-new-eu-data-market-regulations-coherent-and-efficient (“Technical restrictions on access to and re-use of data may result in failures in data markets and data-driven services markets.”); Valéria Faure-Muntian, Competitive Dysfunction: Why Competition Law Is Failing in a Digital World, The Forum Network (24 Feb. 2021), https://www.oecd-forum.org/posts/competitive-dysfunction-why-competition-law-is-failing-in-a-digital-world.

[52] See Rana Foroohar, The Great US-Europe Antitrust Divide, Financial Times (5 Feb. 2024), https://www.ft.com/content/065a2f93-dc1e-410c-ba9d-73c930cedc14.

[53] See, e.g., Press Release, supra note 50.

[54] See infra, Section III.B. Commentators have also made similar claims; see, e.g., Ganesh Sitaram & Tejas N. Narechania, It’s Time for the Government to Regulate AI. Here’s How, Politico (15 Jan. 2024), https://www.politico.com/news/magazine/2024/01/15/sitaraman-artificial-intelligence-regulation-00134873 (“All that cloud computing power is used to train foundation models by having them “learn” from incomprehensibly huge quantities of data. Unsurprisingly, the entities that own these massive computing resources are also the companies that dominate model development. Google has Bard, Meta has LLaMa. Amazon recently invested $4 billion into one of OpenAI’s leading competitors, Anthropic. And Microsoft has a 49 percent ownership stake in OpenAI—giving it extraordinary influence, as the recent board struggles over Sam Altman’s role as CEO showed.”).

[55] Press Release, supra note 50.

[56] Comment of U.S. Federal Trade Commission to the U.S. Copyright Office, Artificial Intelligence and Copyright, Docket No. 2023-6 (30 Oct. 2023), at 4, https://www.ftc.gov/legal-library/browse/advocacy-filings/comment-federal-trade-commission-artificial-intelligence-copyright (emphasis added).

[57] Jonathan Kanter, Remarks at the Promoting Competition in AI Conference (30 May 2024), https://youtu.be/yh–1AGf3aU?t=424.

[58] Karin Matussek, AI Will Fuel Antitrust Fires, Big Tech’s German Nemesis Warns, Bloomberg (26 Jun. 2024), https://www.bloomberg.com/news/articles/2024-06-26/ai-will-fuel-antitrust-fires-big-tech-s-german-nemesis-warns?srnd=technology-vp.

[59] Id.

[60] See, e.g., Joe Caserta, Holger Harreis, Kayvaun Rowshankish, Nikhil Srinidhi, & Asin Tavakoli, The Data Dividend: Fueling Generative AI, McKinsey Digital (15 Sep. 2023), https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-data-dividend-fueling-generative-ai (“Your data and its underlying foundations are the determining factors to what’s possible with generative AI.”).

[61] See, e.g., Tim Keary, Google DeepMind’s Achievements and Breakthroughs in AI Research, Techopedia (11 Aug. 2023), https://www.techopedia.com/google-deepminds-achievements-and-breakthroughs-in-ai-research; see, e.g., Will Douglas Heaven, Google DeepMind Used a Large Language Model to Solve an Unsolved Math Problem, MIT Technology Review (14 Dec. 2023), https://www.technologyreview.com/2023/12/14/1085318/google-deepmind-large-language-model-solve-unsolvable-math-problem-cap-set; see also, A Decade of Advancing the State-of-the-Art in AI Through Open Research, Meta (30 Nov. 2023), https://about.fb.com/news/2023/11/decade-of-advancing-ai-through-open-research; see also, 200 Languages Within a Single AI Model: A Breakthrough in High-Quality Machine Translation, Meta, https://ai.meta.com/blog/nllb-200-high-quality-machine-translation (last visited 18 Jan. 2023).

[62] See, e.g., Jennifer Allen, 10 Years of Siri: The History of Apple’s Voice Assistant, Tech Radar (4 Oct. 2021), https://www.techradar.com/news/siri-10-year-anniversary; see also Evan Selleck, How Apple Is Already Using Machine Learning and AI in iOS, Apple Insider (20 Nov. 2023), https://appleinsider.com/articles/23/09/02/how-apple-is-already-using-machine-learning-and-ai-in-ios; see also, Kathleen Walch, The Twenty Year History Of AI At Amazon, Forbes (19 Jul. 2019), https://www.forbes.com/sites/cognitiveworld/2019/07/19/the-twenty-year-history-of-ai-at-amazon.

[63] See infra Section III.C.

[64] See, e.g., Cédric Argenton & Jens Prüfer, Search Engine Competition with Network Externalities, 8 J. Comp. L. & Econ. 73, 74 (2012).

[65] John M. Yun, The Role of Big Data in Antitrust, The GAI Report on the Digital Economy 220, 233 (2020), https://gaidigitalreport.com/2020/08/25/big-data-and-barriers-to-entry/#_ftnref50; see also, e.g., Robert Wayne Gregory, Ola Henfridsson, Evgeny Kaganer, & Harris Kyriakou, The Role of Artificial Intelligence and Data Network Effects for Creating User Value, 46 Acad. of Mgmt. Rev. 534 (2020) (final pre-print version at 4), http://wrap.warwick.ac.uk/134220 (“A platform exhibits data network effects if, the more that the platform learns from the data it collects on users, the more valuable the platform becomes to each user.”); Karl Schmedders, José Parra-Moyano, & Michael Wade, Why Data Aggregation Laws Could be the Answer to Big Tech Dominance, Silicon Republic (6 Feb. 2024), https://www.siliconrepublic.com/enterprise/data-ai-aggregation-laws-regulation-big-tech-dominance-competition-antitrust-imd.

[66] Nathan Newman, Search, Antitrust, and the Economics of the Control of User Data, 31 Yale J. Reg. 401, 409 (2014) (emphasis added); see also id. at 420 & 423 (“While there are a number of network effects that come into play with Google, [“its intimate knowledge of its users contained in its vast databases of user personal data”] is likely the most important one in terms of entrenching the company’s monopoly in search advertising…. Google’s overwhelming control of user data… might make its dominance nearly unchallengeable.”).

[67] See also Yun, supra note 65 at 229 (“[I]nvestments in big data can create competitive distance between a firm and its rivals, including potential entrants, but this distance is the result of a competitive desire to improve one’s product.”).

^[68] See Manne & Auer, Antitrust Dystopia and Antitrust Nostalgia, supra note 42.

[69] Andrei Hagiu & Julian Wright, Data-Enabled Learning, Network Effects, and Competitive Advantage, 54 RAND J. Econ. 638 (2023).

^[70] Id. at 639. The authors conclude that “Data-enabled learning would seem to give incumbent firms a competitive advantage. But how strong is this advantage and how does it differ from that obtained from more traditional mechanisms… .”

^[71] Id.

^[72] Bruno Jullien & Wilfried Sand-Zantman, The Economics of Platforms: A Theory Guide for Competition Policy, 54 Info. Econ. & Pol’y 10080, 101031 (2021).

^[73] Daniele Condorelli & Jorge Padilla, Harnessing Platform Envelopment in the Digital World, 16 J. Comp. L. & Pol’y 143, 167 (2020).

[74] See Hagiu & Wright, supra note 6969.

^[75] For a summary of these limitations, see generally Catherine Tucker, Network Effects and Market Power, supra note 41; see also Manne & Auer, supra note 42, at 1330.

[76] See Jason Furman, Diane Coyle, Amelia Fletcher, Derek McAuley, & Philip Marsden (Dig. Competition Expert Panel), Unlocking Digital Competition (2019) at 32-35 (“Furman Report”), available at https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/785547/unlocking_digital_competition_furman_review_web.pdf.

[77] Id. at 34.

[78] Id. at 35. To its credit, it should be noted, the Furman Report does counsel caution before mandating access to data as a remedy to promote competition. See id. at 75. That said, the Furman Report maintains that such a remedy should remain on the table because “the evidence suggests that large data holdings are at the heart of the potential for some platform markets to be dominated by single players and for that dominance to be entrenched in a way that lessens the potential for competition for the market.” Id. The evidence, however, does not show this.

^[79] Case COMP/M.9660 — Google/Fitbit, Commission Decision (17 Dec. 2020) (Summary at O.J. (C 194) 7), available at https://ec.europa.eu/competition/mergers/cases1/202120/m9660_3314_3.pdf, at 455.

^[80] Id. at 896.

[81] See Natasha Lomas, EU Checking if Microsoft’s OpenAI Investment Falls Under Merger Rules, TechCrunch (9 Jan. 2024), https://techcrunch.com/2024/01/09/openai-microsoft-eu-merger-rules.

^[82] Amended Complaint at 11, Meta/Zuckerberg/Within, Fed. Trade Comm’n. (2022) (No. 605837), available at https://www.ftc.gov/system/files/ftc_gov/pdf/D09411%20-%20AMENDED%20COMPLAINT%20FILED%20BY%20COUNSEL%20SUPPORTING%20THE%20COMPLAINT%20-%20PUBLIC%20%281%29_0.pdf.

[83] Amended Complaint (D.D.C), supra note 49, at ¶37.

[84] Amended Complaint (E.D. Va), id., at ¶8.

^[85] Merger Guidelines, US Dep’t of Justice & Fed. Trade Comm’n (2023) at 25, available at https://www.ftc.gov/system/files/ftc_gov/pdf/2023_merger_guidelines_final_12.18.2023.pdf.

[86] Merger Assessment Guidelines, Competition and Mkts. Auth (2021) at ¶7.19(e), available at https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1051823/MAGs_for_publication_2021_–_.pdf.

[87] Furman Report, supra note 76, at ¶4.

[88] See, e.g., Chris Westfall, New Research Shows ChatGPT Reigns Supreme in AI Tool Sector, Forbes (16 Nov. 2023), https://www.forbes.com/sites/chriswestfall/2023/11/16/new-research-shows-chatgpt-reigns-supreme-in-ai-tool-sector/?sh=7de5de250e9c; Sujan Sarkar, AI Industry Analysis: 50 Most Visited AI Tools and Their 24B+ Traffic Behavior, Writerbuddy (last visited 15 Jul. 2024), https://writerbuddy.ai/blog/ai-industry-analysis.

^[89] See Krystal Hu, ChatGPT Sets Record for Fastest-Growing User Base, Reuters (2 Feb. 2023), https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01; Google: The AI Race Is On, App Economy Insights (7 Feb. 2023), https://www.appeconomyinsights.com/p/google-the-ai-race-is-on.

^[90] See Google Trends, https://trends.google.com/trends/explore?date=today%205-y&q=%2Fg%2F11khcfz0y2,%2Fg%2F11ts49p01g&hl=en (last visited 12 Jan. 2024) and https://trends.google.com/trends/explore?date=today%205-y&geo=US&q=%2Fg%2F11khcfz0y2,%2Fg%2F11ts49p01g&hl=en (last visited 12 Jan. 2024).

^[91] See David F. Carr, As ChatGPT Growth Flattened in May, Google Bard Rose 187%, Similarweb Blog (5 Jun. 2023), https://www.similarweb.com/blog/insights/ai-news/chatgpt-bard.

^[92] See Press Release, Introducing New AI Experiences Across Our Family of Apps and Devices, Meta (27 Sep. 2023), https://about.fb.com/news/2023/09/introducing-ai-powered-assistants-characters-and-creative-tools; Sundar Pichai, An Important Next Step on Our AI Journey, Google Keyword Blog (6 Feb. 2023), https://blog.google/technology/ai/bard-google-ai-search-updates.

^[93] See Ion Prodan, 14 Million Users: Midjourney’s Statistical Success, Yon (19 Aug. 2023), https://yon.fun/midjourney-statistics; see also Andrew Wilson, Midjourney Statistics: Users, Polls, & Growth [Oct 2023], ApproachableAI (Oct. 13, 2023), https://approachableai.com/midjourney-statistics.

^[94] See Hema Budaraju, New Ways to Get Inspired with Generative AI in Search, Google Keyword Blog (12 Oct. 2023), https://blog.google/products/search/google-search-generative-ai-october-update; Imagine with Meta AI, Meta (last visited 12 Jan. 2024), https://imagine.meta.com.

^[95] Catherine Tucker, Digital Data, Platforms and the Usual [Antitrust] Suspects: Network Effects, Switching Costs, Essential Facility, 54 Rev. Indus. Org. 683, 686 (2019).

^[96] Manne & Auer, supra note 42, at 1345.

[97] See, e.g., Stefanie Koperniak, Artificial Data Give the Same Results as Real Data—Without Compromising Privacy, MIT News (3 Mar. 2017), https://news.mit.edu/2017/artificial-data-give-same-results-as-real-data-0303 (“[Authors] describe a machine learning system that automatically creates synthetic data—with the goal of enabling data science efforts that, due to a lack of access to real data, may have otherwise not left the ground. While the use of authentic data can cause significant privacy concerns, this synthetic data is completely different from that produced by real users—but can still be used to develop and test data science algorithms and models.”).

[98] See, e.g., Rachel Gordon, Synthetic Imagery Sets New Bar in AI Training Efficiency, MIT News (20 Nov. 2023), https://news.mit.edu/2023/synthetic-imagery-sets-new-bar-ai-training-efficiency-1120 (“By using synthetic images to train machine learning models, a team of scientists recently surpassed results obtained from traditional ‘real-image’ training methods.).

[99] Thibault Schrepel & Alex ‘Sandy’ Pentland, Competition Between AI Foundation Models: Dynamics and Policy Recommendations, MIT Connection Science Working Paper (Jun. 2023), at 8.

[100] Igor Susmelj, Optimizing Generative AI: The Role of Data Curation, Lightly (last visited 15 Jan. 2024), https://www.lightly.ai/post/optimizing-generative-ai-the-role-of-data-curation.

[101] See, e.g., Xiaoliang Dai, et al., Emu: Enhancing Image Generation Models Using Photogenic Needles in a Haystack, ArXiv (27 Sep. 2023) at 1, https://ar5iv.labs.arxiv.org/html/2309.15807 (“[S]upervised fine-tuning with a set of surprisingly small but extremely visually appealing images can significantly improve the generation quality.”); see also, Hu Xu, et al., Demystifying CLIP Data, ArXiv (28 Sep. 2023), https://arxiv.org/abs/2309.16671.

[102] Lauren Leffer, New Training Method Helps AI Generalize Like People Do, Sci. Am. (26 Oct. 2023), https://www.scientificamerican.com/article/new-training-method-helps-ai-generalize-like-people-do (discussing Brendan M. Lake & Marco Baroni, Human-Like Systematic Generalization Through a Meta-Learning Neural Network, 623 Nature 115 (2023)).

^[103] Timothy B. Lee, The Real Research Behind the Wild Rumors about OpenAI’s Q* Project, Ars Technica (8 Dec. 2023), https://arstechnica.com/ai/2023/12/the-real-research-behind-the-wild-rumors-about-openais-q-project.

^[104] Id.; see also GSM8K, Papers with Code (last visited 18 Jan. 2023), https://paperswithcode.com/dataset/gsm8k; MATH Dataset, GitHub (last visited 18 Jan. 2024), https://github.com/hendrycks/math.

^[105] See Lee, supra note 103.

[106] Geoffrey Manne & Ben Sperry, Debunking the Myth of a Data Barrier to Entry for Online Services, Truth on the Market (26 Mar. 2015), https://truthonthemarket.com/2015/03/26/debunking-the-myth-of-a-data-barrier-to-entry-for-online-services (citing Andres v. Lerner, The Role of ‘Big Data’ in Online Platform Competition (26 Aug. 2014), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2482780.).

[107] See Catherine Tucker, Digital Data as an Essential Facility: Control, CPI Antitrust Chron. (Feb. 2020), at 11 (“[U]ltimately the value of data is not the raw manifestation of the data itself, but the ability of a firm to use this data as an input to insight.”).

[108] Or, as John Yun put it, data is only a small component of digital firms’ production function. See Yun, supra note 65, at 235 (“Second, while no one would seriously dispute that having more data is better than having less, the idea of a data-driven network effect is focused too narrowly on a single factor improving quality. As mentioned in supra Section I.A, there are a variety of factors that enter a firm’s production function to improve quality.”).

^[109] Luxia Le, The Real Reason Windows Phone Failed Spectacularly, History–Computer (8 Aug. 2023), https://history-computer.com/the-real-reason-windows-phone-failed-spectacularly.

[110] Introducing the GPT Store, Open AI (10 Jan. 2024), https://openai.com/blog/introducing-the-gpt-store.

^[111] See Michael Schade, How ChatGPT and Our Language Models Are Developed, OpenAI, https://help.openai.com/en/articles/7842364-how-chatgpt-and-our-language-models-are-developed; Sreejani Bhattacharyya, Interesting Innovations from OpenAI in 2021, AIM (1 Jan. 2022), https://analyticsindiamag.com/interesting-innovations-from-openai-in-2021; Danny Hernadez & Tom B. Brown, Measuring the Algorithmic Efficiency of Neural Networks, ArXiv (8 May 2020), https://arxiv.org/abs/2005.04305.

[112] See Yun, supra note 65 at 235 (“Even if data is primarily responsible for a platform’s quality improvements, these improvements do not simply materialize with the presence of more data—which differentiates the idea of data-driven network effects from direct network effects. A firm needs to intentionally transform raw, collected data into something that provides analytical insights. This transformation involves costs including those associated with data storage, organization, and analytics, which moves the idea of collecting more data away from a strict network effect to more of a ‘data opportunity.’”).

[113] Josh Sisco, POLITICO PRO Q&A: Exit Interview with DOJ Chief Antitrust Economist Susan Athey, Politico Pro (2 Jul. 2024), https://subscriber.politicopro.com/article/2024/07/politico-pro-q-a-exit-interview-with-doj-chief-antitrust-economist-susan-athey-00166281.

[114] Belle Lin, Open-Source Companies Are Sharing Their AI Free. Can They Crack OpenAI’s Dominance?, Wall St. J. (21 Mar. 2024), https://www.wsj.com/articles/open-source-companies-are-sharing-their-ai-free-can-they-crack-openais-dominance-26149e9c.

[115] Lerner, supra note 106, at 4-5 (emphasis added).

[116] See Clayton M. Christensen, The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail (2013).

[117] See David J. Teece, Dynamic Capabilities and Strategic Management: Organizing for Innovation and Growth (2009).

[118] Antitrust merger enforcement has long assumed that horizontal mergers are more likely to cause problems for consumers than vertical mergers. See Geoffrey A. Manne, Dirk Auer, Brian Albrecht, Eric Fruits, Daniel J. Gilman, & Lazar Radic, Comments of the International Center for Law and Economics on the FTC & DOJ Draft Merger Guidelines, International Center for Law & Economics (18 Sep. 2023), https://laweconcenter.org/resources/comments-of-the-international-center-for-law-and-economics-on-the-ftc-doj-draft-merger-guidelines.

[119] See Hagiu & Wright, supra note 69, at 69 (“We use our dynamic framework to explore how data sharing works: we find that it in-creases consumer surplus when one firm is sufficiently far ahead of the other by making the laggard more competitive, but it decreases consumer surplus when the firms are sufficiently evenly matched by making firms compete less aggressively, which in our model means subsidizing consumers less.”); see also Lerner, supra note 106.

[120] See, e.g., Hagiu & Wright, id. (“We also use our model to highlight an unintended consequence of privacy policies. If such policies reduce the rate at which firms can extract useful data from consumers, they will tend to increase the incumbent’s competitive advantage, reflecting that the entrant has more scope for new learning and so is affected more by such a policy.”); Jian Jia, Ginger Zhe Jin, & Liad Wagman, The Short-Run Effects of the General Data Protection Regulation on Technology Venture Investment, 40 Marketing Sci. 593 (2021) (finding GDPR reduced investment in new and emerging technology firms, particularly in data-related ventures); James Campbell, Avi Goldfarb, & Catherine Tucker, Privacy Regulation and Market Structure, 24 J. Econ. & Mgmt. Strat. 47 (2015) (“Consequently, rather than increasing competition, the nature of transaction costs implied by privacy regulation suggests that privacy regulation may be anti-competitive.”).