David Sacks Rejects Antitrust Immunity for AI Labs Prioritizing Safety

David Sacks, chair of the President’s Council of Advisors on Science & Technology, has stated that AI laboratories must prioritize safety independently without seeking antitrust immunity from the government. His position signals a move toward a market-driven approach to AI risk, rejecting the idea that the state should provide legal protection for companies that coordinate to slow down development.

Sacks argues that leading labs, specifically OpenAI and Anthropic, already effectively function as a duopoly when measured by model capability, revenue growth, and market share. Because of this dominant position, Sacks believes these companies should not receive government "cover" to collectively pace their development. Instead, he maintains that if these firms wish to slow their progress for safety reasons, they should do so independently.

The Push for Coordinated Pacing

This stance comes as frontier AI labs have expressed concerns that intense competitive pressure forces developers to cut corners on safety. To mitigate this "race to the bottom," some industry leaders have suggested a coordinated pacing of frontier model releases. This proposal has sparked a legal debate over whether such coordination would violate antitrust laws, leading some labs to seek government immunity to ensure that safety-first development does not result in prosecution for collusion.

A Shift Toward Product Liability

Rather than establishing a government-led regulatory approval regime, Sacks advocates for a system rooted in product liability to discipline AI development. Under this framework, the responsibility for safety rests with the developers, who would be held legally and financially accountable for the harms caused by their products. This approach shifts the burden of risk from a preemptive bureaucratic process to a retrospective legal one, where the threat of litigation serves as the primary incentive for rigorous safety testing.

Industry Implications

As a top adviser to the U.S. government on science and technology, Sacks' position suggests a significant policy shift. By opposing legal protections for coordinated slowdowns, the administration may increase competitive pressure on the industry's leaders, forcing them to balance rapid innovation with safety without the safety net of antitrust exemptions. This approach prioritizes market competition over the industry's desire for a managed, collective transition to safer AI.

Policy Outlook

While Sacks has clarified his preference for liability over regulation, it remains to be seen how this will translate into formal policy or legislation. It is currently unclear whether the administration will pursue new product liability laws specifically tailored for AI or rely on existing tort law to manage the risks of frontier models. The outcome will determine whether the AI industry moves toward a regime of strict corporate accountability or continues to lobby for state-sanctioned coordination.