A former Google DeepMind researcher has left the company with a serious warning about AI. Bilal Chughtai worked on AGI safety and alignment. He says advanced AI could one day threaten humanity. He left DeepMind in July 2026.
He recently explained his concerns on X. Chughtai said his time at the company made him more and more uneasy about where AI is heading. He wants AI companies to slow down. He also wants them to improve safety work. And he wants them to work together, instead of racing to build stronger systems.
Chughtai is not alone. His comments follow similar remarks from Jacob Coxon, a former Anthropic researcher who also resigned recently. Coxon said some people building AI truly believe it could cause deep harm within ten years. After that, Anthropic scientist Evan Hubinger said he agreed. He put the odds of AI wiping out humanity within a decade at more than ten percent.
These warnings point to one core problem: alignment. This means making sure AI systems follow human goals and stay under human control. Right now, many researchers feel the tools to guarantee this are not keeping pace with how fast AI is growing.
Also Read: Silicon Valley Split Over AI Safety Warnings: What's Really Going On
Big names in AI are now weighing in on this debate. Anthropic CEO Dario Amodei wants the pace of AI development to slow down. OpenAI CEO Sam Altman has talked about the need for more safety cooperation. Meta CEO Mark Zuckerberg sees things differently. He says AI companies already have enough reason to manage safety on their own.
The main question is still the same. How fast should AI grow, and can safety keep up? Chughtai's resignation adds one more voice to this debate. More researchers keep raising concerns about the risks of powerful AI.