Trump's AI Safety Deal Is Just a Promise, Not a Law
Six AI giants signed a pledge — nothing actually forces them to keep it.
Six of the biggest AI companies — Google, OpenAI, Meta, Anthropic, xAI and chipmaker Nvidia — sat down to lunch with President Trump on Tuesday and signed something called 'The White House Accord on Super Intelligence.' Trump described it as 'tremendous self-regulation,' which is a fancy way of saying the companies wrote their own rules and agreed to follow them. This is a handshake, not a law. Nothing in the document is legally binding, and there are no penalties written in.
Here's what they actually promised. First, they'll build strong internal checks to watch whether their AI models behave or go off the rails. Second, they'll give an internal team the power to spot problems and fix them. Third, they'll hire an outside monitor allowed to independently verify all that work. Fourth, their boards of directors will set up a committee to receive reports on it all. In plain terms: they're promising to watch themselves, and to hire someone to watch them watch themselves.
Why should you care? AI now drafts your emails, answers your kids' homework questions, screens your job applications and handles your bank's customer service chat. If one of these systems leaks your data, gives dangerous advice, or gets hijacked by hackers, this accord offers you nothing to point at. No fines. No deadlines. The one real threat is that publicly promising something and not doing it can be treated as a deceptive business practice under the FTC Act — but the usual punishment for that is simply ordering a company not to lie again.
The catch is that Washington isn't actually stepping back. The same week, news broke that the Federal Trade Commission plans to investigate several AI companies over consumer protection concerns. These same companies have also asked for an antitrust exemption so they can coordinate on safety without legal risk — experts say they don't need one. And this isn't even new: the UK and South Korea collected similar voluntary pledges in early 2025. Voluntary promises, it turns out, are easy to sign and easy to forget.
- Six AI giants signed a voluntary safety pledge at the White House — there are no fines or penalties if they break it
- The only real consequence is that the FTC could call it 'deceptive' if they lie, and the usual fix is just a promise to stop lying
- Days later, the FTC said it plans to investigate AI companies over consumer protection issues — so the rules are still being written
Why It Matters
The AI tools you already use stay largely unregulated, governed by promises instead of enforceable laws.