Trump’s Former AI Czar Tells Anthropic and OpenAI to Slow Down and Stop Asking for Waivers
David Sacks, a former White House AI and crypto czar, has advised Anthropic and OpenAI to pace the frontier of AI development but to avoid seeking antitrust waivers or approval regimes. Instead, he argues, product liability and market pressure are sufficient to discipline them.
Sacks, who chairs the President’s Council of Advisors on Science and Technology, made these comments in response to requests from the two AI companies, as reported by ANI. He emphasizes that while he supports their pursuit of technological advancement, he disagrees with their demand for special treatment.
In his letter, Sacks states, "The easiest way not to build superintelligence is for you to agree not to build it." He objects to their proposal for an approval process, suggesting that it amounts to "blackmail."
His concerns echo those raised by an Anthropic researcher who resigned in September, citing fears about the potential risks of AI. Sacks also questions the independence of METR, Anthropic’s evaluation provider, due to its connections with the company’s investors and staff.
Europe has already taken a stand on this matter. In February 2025, the European Commission withdrew its AI Liability Directive, opting instead for strict product liability rules that come into effect on December 9th. These rules treat standalone software and AI systems as products, holding companies liable for defects that lead to cyberattacks or other unsafe behaviors.
Sacks proposes an experiment, drawing inspiration from Europe’s approach, to test an alternative discipline for AI development.