Overview
How should antitrust law adapt when the tools that animate markets are themselves changing at unprecedented speed? Over the next two months, our blogs will explore how artificial intelligence (AI) is rewriting the rules of price fixing, information sharing, market dominance, predatory conduct, and the balance between intellectual property and competition across the United States, the United Kingdom, and Europe.
When the Justice Department filed its complaint against Google for monopolizing the search market in October 2020, artificial intelligence appeared only once in the entire document. Tucked away in paragraph 44 on page 15, the DOJ briefly noted, in subpoint 4, that search services could be delivered through "artificial intelligence software (voice assistance) accessed by a button or voice command."
When Judge Amit Mehta wrote his opinion finding Google liable in August 2024, he dedicated four pages to Google's incorporation of AI into its search processes as part of his findings of fact. However, he concluded, "despite recent advances, AI has not supplanted the traditional ingredients that define general search ... and it is not likely to do so anytime soon."
Just over a year later, when Judge Mehta issued his remedies opinion, AI had become so central to Google's business model that it became a core consideration as Judge Mehta structured his order on relief. He dedicated the first section of his findings of fact, which spanned 30 pages, entirely to generative artificial intelligence (GenAI). In rejecting most of the government's more aggressive proposed remedies, including a prohibition on the kind of preferential deals that led to Google being found liable in the first place, the judge observed that “[t]he emergence of [generative artificial intelligence technology] changed the course of this case,” as GenAI was now “front and center as a nascent competitive threat” to general search engines. As a consequence, Judge Mehta's principal remedy was to require Google to share proprietary search data, which he justified by the usefulness of that data for grounding AI products and integrating AI into search.
So too in the digital market regulation regimes overseen by the European Commission and the UK’s Competition and Markets Authority, the use of AI by gatekeepers such as Apple and Google is increasingly regarded as a tool to entrench market power, whether in mobile ecosystems, search and search advertising, and beyond. Antitrust investigations lean on this emerging view in the examination of mergers involving AI companies and when scrutinising the role of developers of AI models (as well as users).
The proliferation of AI technologies over the past several years has disrupted traditional assumptions about monopoly power and unfair competition. Over the next four weeks, our antitrust and AI colleagues will examine the sudden and often counterintuitive changes that artificial intelligence has compelled in the antitrust context.
- AI Agreements: Our series begins by analyzing new theories under which unlawful agreements may be inferred, outside the algorithmic pricing tools that have been the center of recent antitrust cases. When it becomes industry standard for companies to use the same AI technologies, at what point does shared use of that technology support finding a price-fixing, supply-fixing, customer allocation, or information-sharing agreement? And when can this liability include not only market players, but the developers of the AI tools themselves?
- Unilateral Use of AI: Next, our series examines the ways that a company's unilateral uses of AI technology could violate antitrust laws. Dynamic pricing represents one of AI's key advantages, including maximizing customer value. But what happens when that pricing targets customers most likely to be subjects of competition? Is there a point at which AI-driven pricing can become predatory?
- Achieving an AI Monopoly: Our series continues by addressing the biggest issue foreshadowed in the Google search remedies ruling. Courts and government antitrust enforcement officials are apt to be wary when companies with dominant market shares use their comparative advantages, including vast and unique user datasets, to scale development in ways that stifle emerging competition. As Magnificent Seven companies (Alphabet, Amazon, Apple, Meta, Microsoft, NVIDIA, and Tesla) fight for dominance in burgeoning AI markets, when does leveraging their existing market power to gain competitive advantages become anticompetitive?
- Patents for and Licensing of AI: Finally, our series concludes by turning to the intersection of intellectual property and antitrust law. As key technologies become indispensable for AI deployment and development, certain patents will become essential. How will regulators view refusals to deal by owners of frontier models, hardware (such as GPUs), and datasets as AI increasingly becomes a utility underlying nearly every information technology service and product?
This series will illuminate the path ahead for major technology companies whose business models depend on innovation, traditional industries hoping to harness AI's efficiencies, and consumers seeking to respond to the rapidly changing AI environment. Join us as we explore the cutting edge of antitrust law in the age of artificial intelligence.
