The U.S. government is reported to be preparing to issue an executive order on AI safety and will not force cutting-edge models to receive approval

The Trump administration is preparing to order U.S. agencies to work with artificial intelligence companies (AI) to protect networks from AI-driven cyberattacks, but the order will not force cutting-edge models to undergo government approval, people familiar with the matter said.

Sources said that Trump’s draft executive order will transform the existing cybersecurity information sharing mechanism to include artificial intelligence companies and deal with the threats posed by this emerging technology.

They said the adjustments will help find and patch vulnerabilities in federal, state and local networks and U.S. critical infrastructure, while not creating new regulatory requirements for artificial intelligence models.

It is unclear when Trump will sign the measure, and the executive order could still change.

In response to a request for comment, White House officials said discussions about potential executive orders were speculative and that any policy announcements would be made by Trump himself. Spokespersons for major AI developers such as OpenAI and Anthropic did not immediately respond to requests for comment.