1. United States
2. Alaska
3. Letter

The federal government needs to pursue an international AI safety treaty now, before the technology outpaces any single nation's ability to control it. A former Anthropic and OpenAI researcher's viral warning that AI could kill off humanity wasn't fringe alarmism — OpenAI itself disclosed six instances of "unexpected or concerning model behavior" in just the last six months. A bill to create a nationwide AI kill switch just died in the Senate this week. That's not a policy disagreement. That's a failure of basic governance. California is already moving — Gov. Newsom signed an executive order accelerating AI oversight and called on Congress to return from recess and act. But state-level frameworks can't solve a global race to the bottom. If the U.S. doesn't lead on an international agreement, other nations won't wait. The question Newsom posed is the right one: who is actually in control? Right now, the honest answer is nobody. Push for a binding international treaty with real enforcement mechanisms before that window closes.