Trump: AI leaders sign voluntary accord with reviews

Trump's message on Tuesday was strikingly at odds with warnings from some of the very executives who lunched with him. Dario Amodei of Anthropic had called earlier this month for leading firms to slow down development, and Musk and OpenAI CEO Sam Altman have called for letting safety measures catch up as reports of AI agents going rogue mounted. Yet Trump framed the race as a global contest between the United States and China with no second place, saying he did not want to cooperate too much with China because the U.S. is leading. Not everyone in the room pushed back on that confidence: Jensen Huang and Mark Zuckerberg have downplayed runaway-AI concerns and stressed each company's responsibility for its own products.

The accord Trump described is voluntary, and its details were not immediately available, so what actually gets reviewed and how remains unclear. Alongside it, the White House sought to link AI growth to local goodwill, with executives agreeing to ease bipartisan opposition to data centers by funding local schools or reducing energy costs. That nod to local concerns suggests the political test may be less about model safety than about whether communities accept the data centers AI depends on, and whether the promised reviews amount to more than words.