President Trump signed an executive order Monday that sets up voluntary safety reviews for artificial intelligence and creates a new cybersecurity clearinghouse — but explicitly prohibits any mandatory licensing for AI developers. The order is the administration's latest attempt to shape AI policy without imposing the kind of binding rules that tech companies have fought for years.
Voluntary reviews, not mandates
The order directs federal agencies to develop a framework for companies to submit their AI systems for safety testing on a purely optional basis. Developers who choose to participate would get a government seal of approval, but there's no penalty for skipping the process entirely. The White House framed the approach as a way to encourage best practices without stifling innovation.
That's a sharp contrast with proposals from some lawmakers and consumer groups who have called for mandatory safety assessments before powerful AI models can be deployed. The order makes clear that the administration will not go down that road: it explicitly bans any federal agency from requiring a license or permit to develop or use AI technology.
Cybersecurity clearinghouse
Alongside the safety reviews, the order establishes a centralized clearinghouse for cybersecurity information related to AI. The goal is to give companies and government agencies a single place to share threat data, vulnerability reports, and defensive strategies. The clearinghouse will be run by the Cybersecurity and Infrastructure Security Agency, which already handles similar functions for other critical technologies.
The move addresses a growing concern that AI systems — especially large language models and generative tools — can be exploited by hackers or used to automate attacks. By pooling threat intelligence, the administration hopes to stay ahead of emerging risks without slowing down deployment.
Industry reaction and next steps
Tech industry groups have generally praised the voluntary approach, though some privacy advocates argue it doesn't go far enough. The order does not include any new funding for the safety review process or the clearinghouse; agencies are expected to work within existing budgets.
Federal agencies now have 90 days to propose a detailed plan for the voluntary review framework. The cybersecurity clearinghouse is supposed to be operational within six months. Whether companies will actually line up for voluntary reviews — and whether the system can keep pace with the rapid release of new AI models — remains an open question.
