The United States and technology executives have signed a voluntary accord on safeguards for advanced AI systems, according to news roundups published on Wednesday.

The agreement is voluntary rather than legally binding. It commits signatories to safety practices around the development and deployment of powerful AI models, at a time when governments worldwide are racing to set rules for the technology.

The accord lands as AI regulation becomes a major political issue in Washington. The US Senate this week debated — and Democrats blocked — a Republican bill on data center power costs, with both parties framing AI infrastructure as a cost-of-living issue ahead of November’s midterm elections.

Voluntary pacts of this kind have drawn both praise and scepticism. Supporters say they let companies move faster than legislation; critics argue they lack enforcement teeth and can be abandoned when commercial pressure mounts.

Details of which executives and companies signed, and the accord’s specific commitments, were not available in the reports published so far. Further announcements are expected as the administration and the industry spell out what the safeguards will cover in practice.