Trump Moves to Regulate AI with New Executive Order

President Donald Trump is expected to sign a new executive order on artificial intelligence and cybersecurity early this week, according to sources close to the administration. The directive would set up a voluntary system that requires AI developers to share their models with federal officials 90 days before the public release. It also asks companies to give early access to critical infrastructure owners, such as banks, so that potential risks can be checked in advance. The proposal aims to strike a balance between tech industry concerns and the demands of Trump’s most vocal supporters. On one side, activists linked to former advisers like Steve Bannon and political organizer Amy Kremer want strict government testing of the most powerful AI systems. On the other side, figures from the tech world—venture capitalist Marc Andreessen and former adviser David Sacks—oppose mandatory checks, preferring that companies work with the Department of Commerce’s Center for AI Standards and Innovation on a voluntary basis.The debate has been sparked by the release of new, highly capable AI models such as Anthropic’s Mythos and OpenAI’s GPT‑5. 5‑Cyber. Some experts warn that these tools could be used to launch sophisticated cyberattacks, while others argue the risks are exaggerated. The outcome of the administration’s decision could slow the rollout of large language models or force companies to adjust their designs, potentially affecting profits and innovation. Republican politicians traditionally favor limited regulation, but a growing faction of populist supporters is pushing for AI guardrails. They argue that the government should approve “potentially dangerous” systems before deployment, a view reflected in a recent letter to the White House. Tech leaders, who are major donors and public faces of Trump’s campaign, have called for the Commerce Department to lead the response and encourage voluntary cooperation with security experts. The National Security Agency and cybersecurity officials are also involved in setting up monitoring processes for sudden advances in AI capability. While some think that delaying new models might give the United States a short‑term edge, others warn it will not stop adversaries from acquiring the technology. The administration’s approach may ultimately shape how quickly AI innovations enter the market and how well the country protects itself against emerging cyber threats.