Overview
- The White House postponed a planned executive order on pre-release review of so-called frontier AI models this week, leaving a final signing date unclear and pausing a proposed 90-day voluntary review window.
- Officials remain divided over how long reviews should last and which agencies should have authority, with some national-security advisers pushing for longer, FDA-style approvals and industry leaders warning that strict rules could hurt U.S. competitiveness with China.
- In the interim the government is relying on ad hoc arrangements from the Commerce Department’s NIST unit and its CAISI program, which has set up voluntary pre-release checks with Google, Microsoft and other labs but does not create formal federal approval powers.
- Commercial pressure to ship continues: Google unveiled Gemini 3.5 Flash, the Gemini Omni video generator, and plans to add AI agents across Search, Gmail, Docs and YouTube as companies accelerate product rollouts.
- Legal and social flashpoints are compounding the debate — a jury rejected Elon Musk’s suit against OpenAI on statute-of-limitations grounds and students and writers are publicly protesting AI’s impact on jobs and creative work, sharpening political and cultural stakes.