Partnerships between Big Tech and leading artificial intelligence companies, combined with explosive growth in the field, have made AI a top target for antitrust regulators. Most notably, Microsoft has invested more than $10 billion in OpenAI, the developer of ChatGPT, while Google and Amazon are investing billions in Anthropic, another AI leader. And just last week Microsoft inked a $650 million deal with AI startup Inflection.
The Federal Trade Commission (FTC) and the United States Department of Justice (DOJ) share responsibility for policing competition and the two agencies are engaged in a turf war over which will formally investigate and possibly challenge the partnership between OpenAI and Microsoft. A clearance agreement between the two regulators provides guidelines for which antitrust matters are assigned to which agency, but they are not definitive enough to prevent disputes.
The agency that prevails in this dispute will likely get antitrust authority over the AI sector generally. For a number of reasons, the Justice Department is the better choice for that authority.
Earlier this year, in the middle of this battle, the FTC ordered that OpenAI, Anthropic, Microsoft, Google and Amazon provide information regarding the Big Tech investments and their potential effect on AI competition. The orders were issued under Section 6(b) of the Federal Trade Commission Act, which allows the agency to conduct a study of competition in a particular sector.
FTC Chair Lina Khan, a President Biden appointee, has not hidden her interest in leading the regulatory charge against AI. The 6(b) study may well be her way of jockeying for position in the AI turf war with the DOJ.
Nonetheless, the clearance agreement would seem to favor the Justice Department because it specifies that the leading factor in resolving disputes is “expertise in the product or similar products … gained through a substantial antitrust investigation … within the last seven years.” While neither regulator has significant antitrust experience in the AI sector, Justice has substantial experience in the related area of internet search as a result of its lawsuit against Google.
Internet search and artificial intelligence are merging technologies. Search engines use AI, and AI tools like ChatGPT are often used in similar ways to search engines.
Moreover, the Justice Department alleges that Google’s dominance in internet search gives it a leg up in the AI market, an advantage the DOJ will presumably try to counter in any remedy that results from the lawsuit. That would make FTC enforcement of antitrust in AI duplicative, at least with respect to Google.
Further complicating FTC oversight of AI competition is the agency’s lack of jurisdiction, in most circumstances, over nonprofit entities. This is problematic because OpenAI, the leading AI company, is organized as a nonprofit with a for-profit subsidiary. OpenAI would surely use this fact to its advantage if sued by the FTC. The Justice Department has no such limitation.
Despite the strong arguments for putting the DOJ in charge of competition in the AI industry, the turf war and the FTC’s 6(b) study tell us that agency will not back down easily.
The possibility that the FTC will prevail worries many in the AI industry, as well as antitrust moderates. That’s because the FTC under Khan has signaled that it will pursue a very broad interpretation of its powers under Section 5 of the FTC Act’s prohibition of “unfair methods of competition.”
While past administrations have interpreted Section 5 to prohibit little more than what the main antitrust statutes — the Sherman and Clayton acts — make unlawful, the FTC’s 2022 policy statement makes clear that agency sees no such limitation. The Justice Department, on the other hand, would be bound by the antitrust statutes.
Because the tech giants do not have controlling interests in the leading AI companies and because there is fierce competition in the AI industry, it would be difficult for the government to win a pure antitrust case. That would make the broad use of Section 5 very tempting in an AI case brought by the FTC.
The AI revolution arguably depends on large investments from Big Tech. That’s because training AI models requires an astonishing and ever-increasing amount of computing power that would be difficult to afford without the money and cloud computing capabilities provided by companies like Microsoft, Google and Amazon.
Yet that argument may well fall on deaf ears if it has to be made to the FTC, rather than the DOJ, given the “neo-Brandeis” philosophy Khan has brought to the agency. That approach rejects the current economic efficiency and consumer focused state of antitrust law in favor of using antitrust enforcement more broadly to curb corporate power.
It is not clear that Donald Trump, if elected in November, would choose an FTC chair who is markedly more friendly to the tech giants’ AI investments. After all, Trump is a vocal critic of Big Tech.
What is clear is that, regardless of who heads the FTC, the Justice Department’s expertise in internet search, its jurisdiction over nonprofit entities, and its adherence to the Sherman and Clayton acts make it the best choice to oversee competition in the AI industry.
Curt Levey is president of the Committee for Justice. Before law school, he worked in the AI industry and pioneered technology for overcoming the black box nature of AI models.