Read AI oversight order that President Donald Trump said ‘didn't like certain aspects of it’

President Donald Trump abruptly canceled the signing of a sweeping AI executive order last week, just hours before a scheduled Oval Office event. The draft, which had not been finalized, would have mandated federal safety evaluations for advanced AI models from major tech firms, including Google, Meta, OpenAI, and Microsoft, before their public release. The order, titled *Promoting Advanced Artificial Intelligence Innovation and Security*, emphasized protecting U.S. AI dominance while addressing national security risks. It included provisions to prioritize cyber defenses for government systems, safeguard intellectual property from foreign exploitation, and accelerate secure AI deployment. Trump’s decision to shelve the policy followed his high-profile state visit to China, where he reportedly shifted his stance, arguing that the regulations could hinder America’s lead in AI innovation. According to The New York Times, Trump expressed concerns that the oversight requirements would ‘get in the way’ of maintaining U.S. superiority over global competitors. The draft order had outlined a 30-day timeline for agencies like the Department of War, Homeland Security, and the Cybersecurity and Infrastructure Security Agency (CISA) to implement binding operational directives to strengthen cyber defenses and protect federal information systems. The cancellation marks a reversal from earlier plans to enforce stricter AI safeguards, aligning with Trump’s broader focus on deregulation and competitive positioning in technology. The executive order, which was labeled *PREDECISIONAL* and *FOR DISCUSSION PURPOSES ONLY*, had not yet been officially released but reflected efforts to balance innovation with security risks in AI development.