Trump’s former AI czar tells Anthropic and OpenAI to slow down, and to stop asking for anything in return

David Sacks, former US AI czar, advised Anthropic and OpenAI to slow AI development but reject antitrust waivers or special approval regimes, citing existing product liability and market pressures as sufficient discipline. Europe has similarly classified standalone AI software under strict product liability since December 2025, holding companies accountable for defects including cybersecurity risks without needing to prove fault.