TRUMP SIGNS AI EXECUTIVE ORDER: COMPANIES MUST LET GOVERNMENT TEST MODELS 30 DAYS BEFORE RELEASE President Trump signed an executive order on June 2 that reshapes how the most powerful AI models get released in the United States. The order, titled Promoting Advanced Artificial Intelligence Innovation and Security, calls on frontier AI companies to voluntarily submit their new models to the federal government for testing and evaluation up to 30 days before public release. Washington wants a window to assess what these systems can do before they land in the public’s hands. The framework is voluntary, not mandatory, a key concession the administration made after early drafts drew heavy lobbying from the industry. There are no mandatory licensing requirements and no permitting schemes. But the intent is clear. The government wants early access and the ability to flag potential national security risks before the most capable models go wide. The order also establishes an AI cybersecurity clearinghouse to coordinate between government and industry on software vulnerabilities. It directs the Attorney General to prioritize prosecution of anyone using AI to breach systems or steal data. What it does not do is impose the kind of hard regulatory structure the EU spent years building. This is the American approach. Industry cooperation first, legal force as a backstop if it fails. Keywords: Trump AI executive order, AI regulation 2026, frontier AI models, AI oversight, AI national security