Antitrust Confronts Rapidly Changing AI-Driven Market Structures

PYMNTS reports on comments from Emily Chissell and Nitika Bagaria of Keystone, an antitrust and competition economics consultancy, arguing that AI's competitive effects vary by industry and that regulators face a timing mismatch between slow-moving legal processes and fast-moving AI-driven market structures. Chissell tells PYMNTS that competition authorities are "not generally skeptical about the overall importance and transformative effects that AI is going to have on markets," but are more skeptical of "blanket claims that these disruptions and transformations will always necessarily be good for competition." The piece builds on an April 2026 paper in CPI Antitrust Chronicle, co-authored by Chissell, Bagaria, Dr. Andrea Coscelli, and Tega Akati-Udi. Coscelli is Keystone's Europe co-head and led the UK's Competition and Markets Authority as chief executive from 2016 to 2022.

The April paper frames AI's competitive impact through three cost channels: gen AI can lower fixed operating costs (encouraging entry if the savings are broadly accessible), lower marginal costs (benefiting consumers if passed on, enabling predatory pricing if not), and turn R&D into a strategic arms race, drawing on economist John Sutton's theory of endogenous sunk costs to show how AI can raise or lower entry barriers depending on whether product quality is the main axis of competition. As evidence of how fast these effects can move, the paper's authors cite LSEG's share price falling about 12% and Thomson Reuters' falling about 20% within roughly a week of Anthropic's Opus 4.6 launch and new Claude Cowork features, alongside a similar drop for Publicis Groupe - real-time repricing the authors treat as a market-based signal of dynamic competition. The only UK merger review to date that has meaningfully referenced gen AI use cases is the CMA's Getty Images/Shutterstock inquiry.

This isn't only an argument for slower or tougher merger review: Keystone's own framework equally supports clearing certain deals on efficiency grounds, for example when two smaller firms merge specifically to reach the scale needed to compete in an AI-driven R&D race. Companies preparing AI-related deals or defending AI-driven pricing or bundling decisions should expect regulators to demand compelling, granular, industry-specific evidence, not qualitative claims that "AI changes everything," since Chissell and Bagaria argue regulators will scrutinize AI-related arguments from merging parties carefully rather than accept them at face value.

Track the European Commission's ongoing consultation on revised merger guidelines and the CMA's separate call for evidence on merger efficiencies, both of which are actively soliciting input on how to treat AI-driven dynamic competition. Also watch for the next merger decisions, beyond Getty/Shutterstock, that explicitly weigh gen AI cost effects, and for whether agencies adopt event-study methods (like the share-price reactions cited here) as a standard tool for assessing AI-driven disruption.