Trump's New AI Executive Order Criticized for Lacking Enforcement Teeth
Key Takeaways
- ▸The executive order contains no mandatory compliance requirements; AI companies can entirely opt out
- ▸The government explicitly disclaims liability, removing accountability if approved models cause harm
- ▸Implementing agencies have been severely understaffed and underfunded, making 30-day reviews impractical
Summary
The Trump Administration has issued a new AI executive order titled 'Advanced AI Innovation and Security,' intended to give the federal government 30 days to review 'covered frontier models' before public release for cybersecurity vulnerabilities. However, critics argue the order is largely performative and lacks meaningful enforcement mechanisms. The order explicitly exempts companies from mandatory compliance, allows the government to disclaim all liability, and provides no legal recourse for citizens harmed by approved models—making participation entirely voluntary.
The policy appears prompted by Anthropic's Mythos system, which alarmed industry observers about the risks of unchecked frontier AI. Despite stated intentions, implementation faces severe obstacles. The Cybersecurity and Infrastructure Security Agency (CISA), which would conduct most reviews alongside the NSA and Treasury, has been gutted by budget cuts from Trump's Department of Government Efficiency (DOGE). With no clear funding and minimal staff remaining, the practical capacity to evaluate complex AI systems within 30 days is questionable. Critics argue the order amounts to 'security theater'—appearing to address concerns while establishing no enforceable safeguards.
- The order is criticized as performative regulation with no real enforcement mechanisms