THOMSON REUTERS
Westlaw Today
Enter to open, tab to navigate, enter to select
Addressing competition concerns in the developing generative AI industry: key insights from the FTC's analysis
2023 PRINDBRF 0481
• Practitioner Insights Commentaries
• September 18, 2023
(September 18, 2023) - Ketan Bhirud, Drew Mann and Trey Smith of Troutman Pepper discuss the Federal Trade Commission's role in competition enforcement, contextualize the FTC's analysis of the generative AI industry and provide key takeaways for stakeholders to consider during a period of regulatory uncertainty.
Artificial intelligence has the potential to transform our economy, culture, politics, and civic lives. While the promise of intelligent, independent artificial intelligence systems has not yet materialized at scale, generative artificial intelligence (generative AI) is readily available to hundreds of millions of individuals and businesses today.
In a recent blog post, the Federal Trade Commission (FTC) cautioned that generative AI has the potential to be concentrated in a small number of firms, posing a threat to innovation, competition, and consumer welfare.1 As generative AI becomes an increasingly critical tool in society, the FTC warns that market concentration could lead to a small number of firms wielding outsize influence over a significant portion of the economy.
This article discusses the FTC's role in competition enforcement, contextualizes the FTC's analysis of the generative AI industry, and provides key takeaways for stakeholders to consider during a period of regulatory uncertainty.
The FTC's role in competition enforcement
One of the FTC's stated aims is to ensure that American markets are open and competitive. The agency pursues this mandate by applying consumer protection and antitrust laws to emerging technologies such as generative AI. To that end, the agency proactively researches how technological developments in commerce could harm markets and identifies opportunities to develop the law under its authority.
Much of the agency's authority derives from Section 5(a) of the FTC Act, which prohibits unfair or deceptive acts or practices in or affecting commerce, including conduct that violates the Sherman Antitrust Act or the Clayton Act (antitrust laws). Section 6(g) also authorizes the FTC to promulgate, modify, and repeal trade regulation rules that define acts or practices.
Importantly, the FTC has increasingly utilized its dormant Section 6(g) authority to strengthen its antitrust enforcement capabilities in recent years. For example, over the past two years, the FTC revised its prior policy statements regarding the scope of the meaning of "unfair methods of competition" under Section 5.2
The FTC also recently declared the use of non-compete clauses in employment contracts to be unfair under Section 5.3 Finally, the agency is in the process of revising its merger guidelines alongside the DOJ.4 Taken with the FTC's recent blog post, it is clear that the agency is preparing to increase its scrutiny and enforcement activity against generative AI firms.
The FTC's blog post
In its blog post, the FTC analyzed how the industry's structure is inherently susceptible to high degrees of concentration and provided examples of how firms may engage in unfair methods of competition to entrench or exert market power.
The potential for market concentration
According to the FTC, firms in the AI industry are prone to high market concentration for several reasons.
First, because generative AI requires large amounts of high-quality data to "train" its algorithms, incumbent technology companies have a significant head start over potential market entrants. As a result, incumbents can create superior generative AI models, thereby preventing new entrants from entering the industry.
Second, because generative AI requires large amounts of computational resources, new entrants may be forced to rely on incumbents' pre-trained, "open-source" generative AI models, which are cheaper to train. Such reliance could lead to a market where a few incumbents control the highest-quality pre-trained models.
Third, through network and platform effects, the FTC argues that a first mover could have a superior product because its models, having interacted with users for longer periods, may produce a better product, generating superior products than its rivals.
Standing alone, none of these market conditions inherently violate the antitrust laws. To be actionable under Section 5 of the FTC Act, a firm's conduct must be a method of conduct undertaken in the marketplace, as opposed to an inherent marketplace condition (such as high concentration or barriers to entry).5 Thus, the structure of the generative AI industry, by the FTC's own policies, does not inherently create risk for firms.
However, the FTC's blog post indicates that the agency likely views the generative AI industry as highly concentrated, raising several antitrust concerns for firms that operate in this space. For example, "incipient violations," or violations of antitrust laws by firms that have not gained full-fledged monopoly or market power, are actionable under Section 5.6
Additionally, the DOJ and FTC's draft merger guidelines provide, amongst others, that "[m]ergers should not significantly increase concentration in highly concentrated markets" or "entrench or extend a dominant position."7 Thus, the concentrated nature of the generative AI industry could mean that participants' commercial activities will be subject to heightened scrutiny and, in some cases, give rise to Section 5 liability.
Unfair methods of competition
In addition to discussing the generative AI industry's heightened concentration, the FTC's blog post identified several types of anticompetitive conduct that could potentially create Section 5 liability.
i. Talent and innovation
A key concern the FTC identified is the freedom of generative AI workers. Because generative AI requires a specialized workforce with a rare skillset, the FTC posits that companies may be incentivized to lock workers through non-compete clauses, thereby restricting the free flow of ideas.
The FTC's focus on the free movement of talent comes against the backdrop of the Non-Compete Clause Rule it proposed in January. The rule provides that it is an unfair method of competition for an employer to enter into or attempt to enter into a non-compete clause with a worker.
The rule represents a significant expansion of antitrust law because, before its enactment, only two out of seventeen plaintiffs had been able to successfully challenge a non-compete clause between an employer and a worker.8 Thus, firms that attempt to "lock" employees with non-compete clauses may face liability under Section 5.
ii. Restraints on competition and M&A activity
The FTC's blog post also identified several other unfair methods of competition that could take place in the generative AI industry, such as exclusive dealing, discriminatory behavior, tying and bundling, and M&A activity.
According to the FTC, incumbents with a wide range of products and services could engage in exclusive dealing by using their offerings to funnel users towards their own generative AI products, thereby making it difficult for competitors to penetrate the market.
The FTC also identified the possibility of incumbents in the cloud computing space stifling competition by giving discriminatory treatment to themselves and their partners through exclusive dealing agreements.
The agency also identified the potential for anticompetitive abuses through tying and bundling activities, citing the example of an incumbent that links a generative AI application with its existing products to reduce the value of its competitors' standalone generative AI offerings.
Finally, the FTC argued that firms may unfairly use merger and acquisition activity to consolidate market power. Specifically, the FTC highlighted that large firms may attempt to buy up critical applications and cut off rivals' access to core products, buy complementary applications and bundle them together, or buy up nascent rivals.
These restraints on competition could trigger Section 5 liability for several reasons. Facially, if the aforementioned competition methods are found to exceed competition on the merits, a violation of the Sherman Act and the Clayton Act could lie — giving rise to Section 5 liability.9
Moreover, as previously stated, the FTC's ability to prosecute incipient violations also may create Section 5 liability, even absent a violation of the antitrust laws. Exclusive dealing, tying, and bundling arrangements that "have the tendency to ripen into violations of the antitrust laws by virtue of industry conditions and the respondent's position within the industry" are actionable under Section 5.10
Similarly, M&A activity that has "the tendency to ripen into violations of the antitrust laws" is actionable under Section 5. Thus, generative AI firms may face liability even if an antitrust violation under the Sherman or Clayton Acts is yet to materialize.11
Finally, generative AI firms could face Section 5 liability for any restraints on competition that violate the "spirit" of the antitrust laws. This includes de facto exclusive dealing, tying, or bundling arrangements that use market power to entrench that power or impede competition in the same market and a series of M&A activity that tends to bring about antitrust harms but that individually may not have given rise to a violation.12 The takeaway on this point is that conduct that falls outside the scope of the antitrust laws but that, nonetheless, in the aggregate, tends to increase market power is actionable under Section 5 as well.
The upshot
Based on the FTC's analysis, several key implications and takeaways emerge. First, the agency clearly views the generative AI market as highly concentrated. The blog post serves as a call to increase scrutiny and monitoring of market concentration and anticompetitive behavior, meaning that generative AI firms will continue to be subject to increased scrutiny.
Second, the FTC has powerful rulemaking and enforcement tools to pursue perceived antitrust violations, including violations that are yet to materialize and that fall outside the scope of the antitrust statutes. In line with its efforts to develop competition law to adapt to emerging technologies, the FTC will likely continue utilizing its Section 6(g) authority to promulgate more robust enforcement tools as it learns more about how generative AI impacts competition.
Given these considerations, firms in the generative AI space should be proactive and hire experienced counsel to assist with agency outreach and establish guardrails for compliance. Moreover, stakeholders should work together to foster competition, innovation, and consumer welfare while ensuring the development of effective regulatory frameworks that adapt to the evolving generative AI landscape. By doing so, they can unlock the full potential of generative AI while maintaining a fair and competitive marketplace.
Notes
1 Fed. Trade Comm'n, Generative AI Raises Competition Concerns (June 29, 2023).
2 Fed. Trade Comm'n, Policy Statement Regarding the Scope of Unfair Methods of Competition Under Section 5 of the Federal Trade Commission Act (Nov. 10, 2022) ("Policy Statement").
3 Non-Compete Clause Rule, 88 Fed. Reg. 3482 (Jan. 19, 2023) (to be codified at 16 CFR 910) ("Non-Compete Clause Rule").
4 Merger Guidelines (Draft for Public Comment), U.S. Dep't of Justice & Fed. Trade Comm'n (July 19, 2023) ("Proposed Guidelines").
5 Policy Statement at 8.
6 Id. at 12.
7 Proposed Guidelines at 3.
8 Non-Compete Clause Rule at 3496.
9 Policy Statement at 12.
10 Policy Statement at 13.
11 Id. at 12.
12 Id. at 13.
Ketan Bhirud serves as counsel for Troutman Pepper and is a member of the firm's regulatory investigations, strategy and enforcement practice group, specializing in government investigations as well as administrative, litigation and appellate matters. He is a former government appointee at the state and federal level. He is based in Washington, D.C., and can be reached at [email protected]. Drew Mann is a partner in the firm's consumer financial services practice group and has extensive experience in the antitrust space, both from his time at the Federal Trade Commission and in private practice. He is based in Washington, D.C., and can be reached at [email protected]. Trey Smith, an associate in the regulatory investigations, strategy, and enforcement practice group, earned his J.D. at Washington and Lee University School of Law. He specializes in government investigations and litigation, with a focus on emerging technologies. He is based in Richmond and can be reached at [email protected].
)
Ketan Bhirud)
Drew Mann)
Trey Smith| End of Document | © 2024 Thomson Reuters. No claim to original U.S. Government Works. |