In 2023, the Federal Trade Commission (FTC) continually emphasized the anticompetitive risks associated with Artificial Intelligence (AI). Emphasizing a desire to avoid what the FTC saw as underenforcement in the early Internet era, the agency spent 2023 issuing warning calls and lining up investigative tools to increase its antitrust enforcement activity. All signs point to 2024 as a year of more AI enforcement actions and enhanced scrutiny in the AI space. This Legal Backgrounder focuses on an analysis of the Biden Administration and the FTC’s statements regarding AI antitrust enforcement to predict where companies and practitioners can expect the FTC to focus its enforcement efforts in 2024.
What Is Artificial Intelligence?
Artificial Intelligence broadly refers to technology that is capable of pattern recognition in a way that resembles human learning. MIT has described AI as “machine-learning models that can learn to make a prediction based on data.” Many people think of Generative AI when they think of AI. Generative AI is a certain type of AI which MIT defines as a “machine-learning model that is trained to create new data.” A popular example of this technology is ChatGPT, which is a language model that can answer questions, write essays, and effectively respond to prompts in a way that mimics human conversation.1 All types of AI rely on data, or inputs, to train the models and produce an output. These inputs, and the technology underlying the AI models, are largely where the FTC has focused its antitrust scrutiny.
The Biden Administration’s AI Policy
The FTC’s focus on stopping anticompetitive conduct in the AI space directly aligns with the Biden Administration’s priorities as AI technology expands in use. On October 30, 2023, President Biden issued the “Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence.” This Executive Order highlights the Administration’s serious focus on the future of AI technology—and its accompanying risks—and serves as a public call to action for the FTC to step up its enforcement efforts in the AI space.
The Executive Order emphasizes that the United States has been a leader in developing AI technology and that the Biden Administration is focused on ensuring that the United States’ leadership in the AI space continues. But alongside the possibilities inherent in the growth of AI technology are serious public policy concerns which the Executive Order seeks to address. The Executive Order focuses on a range of potential risks accompanying the rise of AI technology, including infringement of privacy rights, national security threats, potential commercial fraud, and discrimination from AI technology algorithms. The Executive Order also establishes the White House Artificial Intelligence Council, which is an executive branch group dedicated to coordinating AI policies among the federal agencies.
Chief among the issues the Executive Order seeks to address is competition—both how to foster it and how to stop unfair methods of competition as AI technology develops. The Biden Administration sees these as interconnected goals—preventing anticompetitive conduct fosters competition, and supporting competition helps prevent anticompetitive conduct from occurring in a market where only a few large companies control inputs and technology.
The Executive Order proposes a two-pronged approach to supporting innovation and competition in the AI space. First, the White House’s Executive Order focuses on ways to spur growth in the AI space by broadening education and access to AI technology to support a workforce ready to engage and innovate with this technology. It also emphasizes the need to broaden access to funding to enable small start-ups to meaningfully compete. The Order makes clear that the Biden Administration wants the United States to remain a hub of innovation and continue to draw students from around the world.
Existing alongside this desire to affirmatively support AI growth is a desire to ensure that anticompetitive behavior does not proliferate in this relatively nascent industry. Indeed, the Executive Order states that fostering innovation “requires stopping unlawful collusion and addressing risks from dominant firms’ use of key assets such as semiconductors, computing power, cloud storage, and data to disadvantage competitors.” The Executive Order repeatedly emphasizes the need for “small” businesses and developers to have a chance to compete and innovate in the AI space, signalling a concern with a situation in which a small number of companies dominate the AI industry. The Executive Order also warns of the potential for AI technology to “lessen market competition.”
The Executive Order further directs agency heads to take concrete actions to promote fair competition, including “addressing risks arising from concentrated control of key inputs, taking steps to stop unlawful collusion and prevent dominant firms from disadvantaging competitors.” In so directing, the Order calls out the Federal Trade Commission “in particular” to decide “whether to exercise the Commission’s existing authorities, including its rulemaking authority . . . to ensure fair competition in the AI marketplace and to ensure that consumers and workers are protected from harms that may be enabled by the use of AI.” The Biden Administration’s Executive Order could not be more clear about the role it sees the FTC playing in regulating the AI industry.
FTC Antitrust Enforcement in AI Markets
Perhaps no agency is more eager to answer President Biden’s call to crack down on AI than the FTC. Even before the White House issued its Executive Order on AI, the FTC publicly committed to combatting perceived threats to competition posed by AI technologies.
In a May 2023 op-ed titled “We Must Regulate A.I. Here’s How,” FTC Chair Lina Khan provided a glimpse into the Commission’s strategy to promote competition in artificial intelligence markets. Khan sees the emerging market for AI technology at an inflection point, much like the beginning of the Web 2.0 era about two decades ago. In Khan’s judgment, companies like Facebook and Google enjoyed initial success by competing on the merits of innovative products, but later locked out competition by employing coercive, exclusionary tactics. The FTC Chair not only blames Big Tech companies themselves for concentrating tremendous market power, but also her predecessors in government for “a broad range of policy choices” that allowed them to do so.
Unsurprisingly, then, Khan believes the FTC and other agencies “have a responsibility to ensure this hard-learned history doesn’t repeat itself.”2 To that end, the FTC has recently redoubled its commitment to vigorous antitrust enforcement in AI markets. Speaking in the heart of Silicon Valley at Stanford’s Institute for Economic Policy Research last month, Chair Khan argued “why open, fair, competitive markets are critical to promoting innovation and unleashing the full potential of emerging technologies.” Khan’s remarks at Stanford also touched on AI issues beyond antitrust like imposter schemes and consumer privacy, but her message at its core echoed a common refrain: “there is no AI exemption from the laws on the books.”
The following sections describe the FTC’s top AI-related antitrust concerns and explain how the Commission plans to exercise its regulatory and enforcement authority to address these purported threats to competition.
Monitoring Bottlenecks: The Inputs of AI
Monopolization is among the FTC’s greatest antitrust concerns, as Chair Khan believes the expansion of AI “risks further locking in the market dominance of large incumbent technology firms” that control the resources smaller companies need to develop their own AI technologies.3 The Commission is particularly concerned about monopolization in markets involving generative AI. In a June 2023 blog post, the FTC’s Bureau of Competition and its Office of Technology argued that the largest tech firms could entrench their dominant positions by coercively controlling “essential building blocks of generative AI.” In the agency’s eyes, AI markets encompass three key inputs that are susceptible to anticompetitive abuse: data, labor, and computational resources.4
Much like the algorithms long used by Google and Facebook, AI models are only as good their underlying data. For this reason, access to quality data can be a substantial barrier to entry for smaller companies trying to compete in AI markets. This is especially true when incumbent firms control platforms that collect massive amounts of data and own proprietary data collection and analysis tools. While “simply having large amounts of data is not unlawful,” the FTC has warned the tech industry against using—and sued tech companies for employing—“unreasonable, unfair, or deceptive” data practices.5 The FTC has also warned that “even [companies] with responsible data collection practices” will be subject to antitrust scrutiny if they coercively control data to exclude new entrants or otherwise impede competition.6
Competing in artificial intelligence markets also requires a uniquely skilled workforce, including engineers and researchers with strong technical abilities and special expertise. Since all companies in the AI sector draw from the same scarce pool of qualified workers, firms may seek to restrain competition by implementing artificial barriers to labor mobility.7 Therefore, the FTC finds it “critical that talented individuals with innovative ideas be permitted to move freely” and is likely to continue challenging non-compete agreements.8 Although its legal authority to do so is questionable, the Commission has even proposed a nationwide ban on non-competes intended to promote competition in both labor markets and product markets.9
Creating and operating AI systems also demands tremendous computational resources, which generally come in two forms. The largest tech companies may be able to purchase and maintain expensive hardware, but smaller firms typically rely on cloud computing services. Since these cloud services are provided by only a handful of firms, the FTC is increasingly concerned about anticompetitive conduct in this area and has solicited public comments on the business practices of cloud computing providers. While open-source AI foundation models may promote competition by reducing the amount of computing power needed to train AI, the FTC is also concerned about “open first, closed later” tactics and the strategic use of open-source models to establish market dominance and lock out competitors.10
Other Anticompetitive Conduct
Beyond monopolization through the potentially coercive control of inputs, the FTC has committed to examine a variety of other allegedly exclusionary practices related to artificial intelligence.
Tying & Bundling
Chair Khan considers IBM’s decision to unbundle its software and hardware products in 1969 as a catalyst for multi-trillion dollar growth in the American software industry.11 Just as the Antitrust Division prompted IBM’s unbundling over five decades ago, the FTC intends to promote innovation in emerging AI markets by investigating attempts to exclude competition by bundling and tying.12 Therefore, dominant tech firms that tie or bundle existing core products with new AI technologies are likely to face antitrust scrutiny from the FTC and should be prepared to show how these sales tactics benefit consumers or are otherwise procompetitive.13
The FTC has identified two scenarios of concern regarding exclusive dealing arrangements. First, an incumbent firm may “use their power in the compute services sector to stifle competition in generative AI by giving discriminatory treatment to themselves and their partners over new entrants,” potentially through “exclusive cloud partnerships.”14 Second, an incumbent may offer “both their own products leveraging generative AI as well as offering APIs allowing other companies to leverage their generative AI capabilities,” the terms of which could insulate the firm from competition.15
The Commission has recently challenged exclusive dealing arrangements in the healthcare industry and may use similar litigation strategies to address exclusive dealing in AI markets. In particular, the FTC is exploring antitrust actions against companies that exploit copyrighted training data.16
Network & Platform Effects
Since generative AI models can be improved by “positive feedback loops…, generative AI products can get better the more people use them.”17 The FTC is concerned that these network effects could allow dominant firms to impede market entry and concentrate their market share.18 Thus, the Commission believes that “legal or policy intervention” may be necessary to defend competitors and consumers against dominant firms’ exploitation of network effects. 19
Likewise, since users of generative AI may become dependent on a dominant platform, the FTC is likely to exercise its enforcement authority to address perceived abuses of platform effects. The FTC is paying close attention to platform effects in cloud services markets because smaller companies that rely on cloud providers for computing resources could be subject to lock in and become susceptible to extortionary data egress fees.20
AI Pricing Tools
As anyone who’s booked a flight or called a rideshare during a surge has experienced, AI tools are used in a variety of industries to set prices. With the expanding implementation of AI price-setting tools—even among brick-and-mortar businesses—the FTC is increasingly monitoring the use of AI “to facilitate collusive behavior that unfairly inflates prices, precisely target price discrimination, or otherwise manipulate output.”21 Therefore, companies that use artificial intelligence to set prices in response to consumer demand should be careful to avoid the use of competitors’ confidential pricing information. To help businesses with dynamic pricing models avoid antitrust liability, FTC has offered a simple rule of thumb: “if it isn’t ok for a guy named Bob to do it, then it probably isn’t ok for an algorithm to do it either.” 22
2024: The Year the FTC Takes on AI
With a presidential election on the horizon in 2024, and a steady drumbeat of statements from the FTC and Biden Administration warning of the anticompetitive risks posed by AI technology, 2024 should prove to be the FTC’s most aggressive yet with respect to the AI industry.
Civil Investigative Demands (CIDs) and AI
On November 21, 2023, the FTC announced that it had approved a resolution which would allow it to issue civil investigative demands (CID) “in nonpublic investigations involving products and services that use or claim to be produced using artificial intelligence (AI) or claim to detect its use.” A CID is a mandatory request for information, testimony, or documents and signals that the issuing agency is investigating the target of the CID.
This step signals that the FTC is preparing for new investigations regarding companies using AI or operating in the AI industry. While a company’s mere receipt of a CID does not mean that any enforcement action will ultimately occur, it does mean that the agency in question is advancing in its investigation and is likely considering enforcement action, which could open the target up to an expensive compliance process, costly and time-consuming litigation, and monetary penalties. In fact, the resolution specifically states that CIDs can be used for investigations “to determine whether Commission action to obtain monetary relief would be in the public interest.” Any company or individual who receives a CID should immediately contact legal counsel to advise on compliance and develop a strategy for next steps.
A Path Forward in 2024: Due Diligence and Legal Counsel
While the FTC has cemented its place as the leading antitrust enforcer in the AI space, it is far from the only federal agency pursuing AI regulation. This spring, Chair Khan and leaders from the CFPB, EEOC, and DOJ released a Joint Statement on Enforcement Efforts Against Discrimination and Bias in Automated Systems. In the statement, the “agencies reiterate[d] their resolve to monitor the development and use of automated systems” and “pledge[d] to vigorously use our collective authorities to protect individuals’ rights regardless of whether legal violations occur through traditional means or advanced technologies.”
As this Legal Backgrounder makes clear, companies operating in the AI space and those relying on AI technology should be prepared for CIDs, new investigations, and highly scrutinized mergers in 2024 and beyond. This increased scrutiny means that companies relying on or considering adopting AI technology should not only do their due diligence on the technology or products they are using but should also seek out counsel to advise on the incumbent risks associated with increased FTC enforcement.
Though AI technology promises to transform many industries, its newness and novelty will not shield it from enforcement actions. It is not yet clear how AI will shape our world in the years to come, but regardless of what happens, the FTC will be at the forefront in combatting any perceived anticompetitive conduct.
- Kevin Roose, How Does ChatGPT Really Work?, The New York Times (Mar. 28, 2023).
- Lina M. Khan, We Must Regulate A.I. Here’s How, The New York Times (May 3, 2023).
- “Dominant firms could use their control over these key inputs to exclude or discriminate against downstream rivals, picking winners and losers in ways that further entrench their dominance.” Khan, supra, note 1.
- Also known as “compute.”
- Staff in the Bureau of Competition & Office of Technology, Generative AI Raises Competition Concerns, Fed. Trade Comm’n (June 29, 2023) (“FTC Blog Post”).
- “Research suggests that incumbent firms may be capturing large amounts of innovative capacity through hoarding talent or acquiring firms for the express purpose of killing competition.” Lina M. Khan, Remarks at the Stanford Institute for Economic Policy Research (Nov. 2, 2023).
- Id. at 4.
- FTC Blog Post, supra, note 5.
- Khan, supra, note 2.
- FTC Blog Post, supra, note 5.
- “The law on tying is changing. Although the Supreme Court has treated some tie-ins as per se illegal in the past, lower courts have started to apply the more flexible ‘rule of reason’ to assess the competitive effects of tied sales. Cases turn on particular factual settings, but the general rule is that tying products raises antitrust questions when it restricts competition without providing benefits to consumers.” Tying the Sale of Two Products, Fed. Trade Comm’n.
- FTC Blog Post, supra, note 5.
- “[C]onduct that may be consistent with the copyright laws nevertheless may violate Section 5. Many large technology firms possess vast financial resources that enable them to indemnify the users of their generative AI tools or obtain exclusive licenses to copyrighted (or otherwise proprietary) training data, potentially further entrenching the market power of these dominant firms.” U.S. Copyright Office, Comment of the United States Federal Trade Commission (Oct. 30, 2023) (“Copyright Office Comment”).
- FTC Blog Post, supra, note 5.
- Copyright Office Comment, supra, note 16 (citing Khan, supra, note 2).
- Maureen K. Ohlhausen, Should We Fear The Things That Go Beep In The Night?, Fed. Trade Comm’n.