# DeepMind exit adds to AI safety alarm

> A former Google DeepMind researcher quit with a stark warning as AI leaders and the White House split over how fast to move.
- Title: DeepMind exit adds to AI safety alarm · New York Post
- Summary: A former Google DeepMind researcher quit with a stark warning as AI leaders and the White House split over how fast to move. The fight is over pace: AI…
- Keywords: bilal chughtai, google deepmind, artificial intelligence, agi safety, ai alignment, bluedot impact, jacob coxon, anthropic, dario amodei, sam altman, elon musk, donald trump
- Source: New York Post — https://nypost.com/2026/09/15/business/google-deepmind-researcher-quits-with-chilling-ai-warning-it-could-kill-us-all
- Published: 2026-09-15T20:11:36+00:00
- Read time: 1 min
- Topics: bilal chughtai, google deepmind, artificial intelligence, agi safety, ai alignment, bluedot impact, jacob coxon, anthropic, dario amodei, sam altman, elon musk, donald trump
## DeepMind researcher quits
Bilal Chughtai left Google DeepMind after working on AGI safety and alignment, saying firsthand exposure to AI progress made him fear where the technology is headed.

> AI has the potential to kill us all
## Progress spooked him
Chughtai said AI has advanced at a staggering pace since 2022, especially as agents grow more autonomous and could soon exceed human ability across domains.

> far exceed human capabilities in every domain
## Alignment remains unsolved
His central fear is misalignment: systems may not deeply want what humans want, and could escape control with catastrophic results.

> Alignment is the problem of preventing this
## He wants a slowdown
Chughtai said AI can still be built safely, but only if companies stop racing ahead and give society time to handle new risks before harm arrives.

> We need to pace AI development
## Other insiders agree
Jacob Coxon, formerly at Anthropic and OpenAI, also quit over extinction fears and accused major AI labs of gambling with human lives.

> gambling with our lives
## CEOs urge restraint
Anthropic CEO Dario Amodei called for slowing the most powerful models. Sam Altman and Elon Musk backed the push, while Amodei warned military incentives make coordination hard.

> we should try
## Trump rejects the warnings
President Trump called AI doomsday fears a hoax and argued the US must keep its lead over China. The White House is considering only limited safeguards, not broad regulation.

> limited safeguards
## Key takeaway

The fight is over pace: AI insiders want slower development until risks are clearer, while the White House fears strict rules could weaken the US lead.