Trump delays AI security executive order

President Donald Trump has delayed signing an executive order that would establish a government process to evaluate artificial intelligence models before they are released to the public. Trump attributed the delay to his dissatisfaction with the specific language of the draft, telling the White House press pool that he disliked certain aspects of the proposal. He expressed concern that the current phrasing could act as a blocker to progress, emphasizing a desire to protect competitive leadership. “We’re leading China, we’re leading everybody, and I don’t want to do anything that’s going to get in the way of that leading,” Trump said.

The anticipated executive order would task the Office of the National Cyber Director—the US government agency responsible for cybersecurity policy—and other federal agencies with developing a mandatory security evaluation process. This process would allow the government to evaluate AI models for security before their release.

According to a report by CNN, one of the key sticking points in the draft language is a proposed requirement for AI companies to share advanced models with the government between 14 and 90 days ahead of their launch. This measure was prompted by concerns over the capabilities of models like Anthropic’s Mythos and OpenAI’s GPT-5.5 Cyber, which can quickly find and exploit security vulnerabilities.

Beyond the policy disagreements, several reports indicate that the delay was also driven by scheduling conflicts. The formal signing event in Washington, D.C. was reportedly postponed because not enough technology chief executives could make it to the capital on short notice.

Why it matters

The proposed executive order would have established a mandatory security evaluation process for AI models before release, a move prompted by concerns over the capabilities of models like Anthropic’s Mythos and OpenAI’s GPT-5.5 Cyber.