Donald Trump Hosts Summit as AI Firms Sign Voluntary Safety Accord

What changed

Donald Trump hosted a White House summit on Tuesday where leaders from OpenAI, Anthropic, Nvidia, Meta, Google and SpaceX signed a voluntary AI-safety accord.

The companies accepted responsibility for securing their own systems, using safeguards, working with independent auditors and quickly fixing problems. Trump also said he would create an AI-safety board, while backing rapid expansion of US data centres.

Why it matters

The agreement puts the companies in charge of policing technology whose risks they are also being urged to slow down. That makes the promised audits and corrections important, but the practical force of the deal is still unclear: its enforcement, legal effect and audit rules were not specified.

Trump’s support for more data centres creates a second pressure point. AI needs enormous computing infrastructure, but communities must live alongside the electricity demand and construction. The report says 65% of surveyed registered voters opposed data centres being built in their areas, turning local resistance into a political problem ahead of the November midterm elections.

The likely test is whether the companies’ voluntary safeguards become visible, enforceable practice, and whether their promised contributions persuade communities to accept more facilities. The oversight board’s membership, authority and timetable will show how much of this is government supervision and how much remains corporate self-regulation.