The Trump administration's AI doomer moment

Q: The Trump administration is drafting an executive order to create an AI working group that would imp

The Trump administration is drafting an executive order to create an AI working group that would implement pre-release government review of frontier models — reversing a policy it killed on day one of Trump's second term.

Q: Anthropic's Mythos model, capable of generating cybersecurity exploits, spooked national security of

Anthropic's Mythos model, capable of generating cybersecurity exploits, spooked national security officials and triggered the policy reversal; the White House now opposes expanding Mythos access beyond ~50 companies.

Q: Google, Microsoft, and xAI have already agreed to submit models for early government review through

Google, Microsoft, and xAI have already agreed to submit models for early government review through the Commerce Department's Center for AI Standards and Innovation (formerly the US AI Safety Institute).

Casey Newton·Wednesday, May 6, 2026·10 min read

AI/ML Regulation Politics

Share𝕏 in

AI Summary

The Trump administration is reversing its anti-regulation stance on AI after Anthropic's new Mythos model demonstrated dangerous cybersecurity capabilities, prompting discussion of a pre-release government review process similar to what Biden had implemented. Google, Microsoft, and xAI have already agreed to give the government early access to their models for review. The shift signals declining influence of accelerationists like David Sacks within the administration and growing recognition of frontier AI risks.

Key Facts

✓The Trump administration is drafting an executive order to create an AI working group that would implement pre-release government review of frontier models — reversing a policy it killed on day one of Trump's second term.

✓Anthropic's Mythos model, capable of generating cybersecurity exploits, spooked national security officials and triggered the policy reversal; the White House now opposes expanding Mythos access beyond ~50 companies.

✓Google, Microsoft, and xAI have already agreed to submit models for early government review through the Commerce Department's Center for AI Standards and Innovation (formerly the US AI Safety Institute).

Author Takes

BullishCasey Newton

Trump administration's AI safety reversal

While the administration should have taken AI safety seriously all along, Newton says he will settle for them taking it seriously now, viewing the reversal as a net positive despite its political motivations.

SkepticalCasey Newton

AI licensing regimes under Trump

Newton warns the administration could use any AI model licensing regime for censorship — blocking 'woke' outputs or pressuring companies into political favors — analogous to Brendan Carr's FCC.

BearishCasey Newton

Trump AI policy prior to reversal

Newton previously called the administration's 'let's see what happens' AI policy 'dangerous negligence,' and views its dismissal of safety concerns as having been proven wrong by Mythos.

More from Casey Newton

The best argument I’ve heard for why AI won't take your job

Platformer launches a new podcast mini-series on AI and jobs, featuring Box CEO Aaron Levie arguing that AI will transform rather than eliminate most

May 13

The week that Meta employees became training data

Meta has implemented invasive monitoring software called MCI on employee computers to capture their every mouse movement, click, and keystroke to trai

Apr 24

Sam Altman’s second thoughts

The newsletter examines escalating violence against AI executives, including a Molotov cocktail attack on Sam Altman's home and shootings at data cent

Apr 14

AI Summary

Key Facts

Author Takes

Related topics

More from Casey Newton