Trump delays AI executive order after tech industry pushback

U.S. President Donald Trump postponed the signing of an executive order regulating advanced artificial intelligence models amid pushback from tech industry figures. The decision followed objections from SpaceX and Tesla CEO Elon Musk, Meta CEO Mark Zuckerberg, and venture capitalist David Sacks, though OpenAI supported the measure. The draft order, shared with industry representatives on May 20, proposed requiring developers to submit frontier AI models—such as Anthropic’s Claude Mythos Preview—for federal review 90 days before public release. It also called for early access to these models for critical infrastructure operators and tasked the Treasury Department with overseeing the program, with support from agencies like the NSA and Cybersecurity and Infrastructure Security Agency. The order defined ‘covered frontier models’ as systems requiring a classified benchmarking process to assess their risks, including cybersecurity threats. Claude Mythos Preview, launched in April, has demonstrated the ability to identify thousands of high-severity software vulnerabilities and chain zero-day flaws into complex cyberattacks. The draft also included provisions for federal agencies to expand AI-driven cybersecurity defenses and fund related initiatives through existing grant programs. Trump reportedly invited key tech leaders—including Anthropic CEO Dario Amodei, Zuckerberg, and OpenAI CEO Sam Altman—to the signing event but was informed they would not attend, contributing to the delay. Separately, the White House has reportedly allowed OpenAI to support state-level AI regulations, despite Trump’s 2025 executive order aiming to limit such laws in favor of a national framework. The stalled order reflects ongoing tensions between regulatory ambitions and industry resistance in shaping AI governance.