Former US DOJ Antitrust Chief: AI Giants Seeking Regulatory Exemptions May Be Engaging in Cartel Collusion

Overview

Jonathan Kanter, former head of the Antitrust Division at the U.S. Department of Justice, recently warned that tech giants in the artificial intelligence sector seeking regulatory exemptions may be engaging in de facto cartel collusion. This statement comes amid fierce debate over AI safety and regulation, following the departure of researchers from major labs including Anthropic and Google DeepMind, who cited concerns about model risks. Meanwhile, CEOs of these companies have called for a slowdown in development and even voluntarily requested antitrust exemptions to coordinate their actions under the guise of "safety."

Legal and Regulatory Prospects

Kanter, now a professor at the University of Washington School of Law and a professor of technology policy at Carnegie Mellon University, stated plainly during his appearance on the Decoder podcast that the current AI industry is like having "invented the car but having no traffic rules." He emphasized that the federal government should remain engaged in regulation, rather than adopting the evasive stance currently taken by the Trump administration.

"Although they claim it is for public safety, this collective action poses significant risks under antitrust law."

Kanter believes that although there is intense market competition among major model vendors and mutual distrust between them, they collectively call for government intervention to establish uniform safety standards. This phenomenon fits the classic game theory scenario of the prisoner's dilemma—enterprises hope to leverage external enforcement power to constrain competitors, thereby locking in their own dominant positions.

Cartel Allegations and Motivation Analysis

Regarding the exemption requests filed by companies such as OpenAI, Anthropic, and Google DeepMind, Kanter offered two interpretations ranging from good faith to cynicism, but reached a consistent conclusion: antitrust exemptions should not be granted.

• Good Faith Interpretation: Companies genuinely worry about unchecked innovation in a regulatory vacuum, fearing that "not seeing the road markers" will lead to accidents.
• Harsh Interpretation: This is a classic case of regulatory capture, aimed at excluding competition from low-cost open-source models from China and other regions by having the government set high compliance thresholds.

Kanter specifically pointed out that this strategy may also serve enterprises' capital operation needs. On the eve of an IPO, tech giants are attempting to persuade the government to build compliance barriers, thereby reducing investor pressure regarding high-risk projects while simultaneously curbing the entry of cheaper, more open alternatives into the market.

Conclusion

U.S. antitrust policy is currently undergoing a comprehensive restart. Even David Sacks, a former government official considered a liberal, retweeted Lina Khan’s view that "no antitrust exemptions are needed." This indicates that regardless of political affiliation, the efforts of AI giants to form a regulatory alliance under the pretext of safety are facing widespread scrutiny from both the legal community and the political sphere. How to balance technological innovation with antitrust enforcement in the future will become a key issue in the U.S.-China tech rivalry.