The chair of the Senate Subcommittee on Antitrust, Competition Policy and Consumer Rights, Amy Klobuchar (D-MN), along with several of her Democratic colleagues, this week sent a letter to the chair of the Federal Trade Commission, Lina Kahn, and the head of the Justice Department’s antitrust division, Jonathan Kanter, to urge the two agencies to investigate whether new AI-driven search tools such as Google's Overview feature, violate antitrust laws.
“Recently, multiple dominant online platforms have introduced new generative AI features that answer user queries by summarizing, or, in some cases, merely regurgitating online content from other sources or platforms,” the letter reads in part. “Although these features may provide partial citations or links to sources, they are often hidden behind tabs or at the bottom of a page where users are unlikely to scroll after already reading an answer… When a generative AI feature answers a query directly, it often forces the content creator—whose content has been relegated to a lower position on the user interface—to compete with content generated from their own work…
[W]e urge the Department of Justice Antitrust Division and the Federal Trade Commission to investigate whether the design of some generative AI features, introduced by already dominant platforms, are a form of exclusionary conduct or an unfair method of competition in violation of the antitrust laws.”
The FTC and DOJ hardly need prompting these days to investigate dominant tech companies. The senators’ letter, in fact, was released the same week testimony began in federal court in the second major antitrust lawsuit DOJ has brought against Google this year. In July, U.S. District Court Judge Amit Mehta ruled that Google acted illegally to maintain its monopoly over internet search and is now considering remedies in that case. In the latest case, DOJ is targeting Google’s actions to maintain an alleged monopoly over the online ad-tech ecosystem.
For its part, the FTC has been a vocal advocate for using antitrust and consumer protection laws to police the conduct of AI companies (see here and here).
Most of the friction between rights owners and AI companies has come from disputes over the use of the use of copyrighted works to train AI models. As of the middle of 2024, at least 32 separate lawsuits had been filed in U.S. courts charging AI companies with copyright infringement over the unlicensed use of protected works in training. But as I’ve argued in previous posts, antitrust might provide rights owners with a more direct, or at least shorter, path to protecting their interests than copyright.
The questions raised by all those copyright lawsuits are likely years away from being fully resolved. Courts move slowly, and the law is genuinely unsettled as to whether AI training actually implicates the exclusive rights of authors. The question is as much a technical one as a legal one, and the technology of generative AI can be daunting to get your head around. Most courts are not overly endowed with technical expertise, which could lead to different outcomes in similar cases, further delaying any final resolution.
In contrast, basic antitrust and consumer protection laws have been on the books for decades. And unlike allegations of copyright infringement, which are mostly left to private plaintiffs except in rare cases of alleged criminal infringement, federal and state governments can bring civil cases against alleged violations of antitrust laws and have extensive investigative powers not often available to private parties.
While antitrust regulators in the past have been shy to enforce the antitrust laws aggressively against digital companies often perceived as responsible for economically valuable innovation, the current DOJ and FTC have shown no such compunction.
The senators’ letter suggests that U.S. policymakers also may be warming to the idea of incorporating antitrust and consumer protection considerations into an eventual regulatory framework for AI.
Taking Down the Ad-Tech Stack
Google’s second major antitrust trial this year, currently underway in federal court in Northern Virginia, is not explicitly about AI. But it’s outcome could well affect how growing concern over Big Tech’s dominance of the AI market could develop.
The Justice Department is suing Google for illegally maintaining a monopoly over all three major segments of the ad-tech ecosystem: the software online publishers use to offer their display advertising space for sale, the network of advertisers seeking bid on that space, and the marketplace both sides use to conduct transactions, from which Google takes a cut.
In order for each side to gain full access to the other, the government alleges, Google can and does compel them to use its marketplace, called an ad-exchange, increasingly Google’s profits and harming competing ad-exchanges. The government wants to force Google to divest its Ad Manager platform used by publishers, which could make it easier for publishers to direct sales through non-Google exchanges.
Should the government prevail, it would be the second time this year that Google is found to be an illegal monopoly. A verdict against Google in the current case could have even greater impact than the ruling in the first case regarding Google’s monopoly in search, however, because of how integral the ad-tech stack is to Google’s overall business, and because Google participates directly in the online advertising market.
When a user conducts a search through Google, Google monetizes the user’s engagement and attention by selling ads against the search results. Publishers have long complained that diverts ad dollars to Google that would otherwise have gone to the publisher of whatever website the user clicked to. They view Google’s display of their content in search results as akin to an unlicensed use of their copyrighted content for which the ad dollars they generate from search traffic are inadequate compensation, and have sought the government’s help to prod Google to compensate publishers directly for displaying their content.
Senator Klobuchar and her colleagues now essentially accuse Google and other AI search engines of diverting an even larger share of the ad pie to themselves by providing users with summaries that lessen the need to click through to the sources of the information — if the sources are even identified. And they’re framing as a matter of antitrust law rather than copyright.