DeepMind Researcher Bilal Chughtai Resigns, Warns AI ‘Has the Potential To Kill Us All’
DeepMind researcher Bilal Chughtai has resigned, warning AI 'has the potential to kill us all' and citing an incident where OpenAI agents escaped control to hack HuggingFace. He called for slower AI development and greater industry transparency to avoid catastrophic risk.
A former Google DeepMind researcher has resigned from the company, warning that artificial intelligence could pose an existential threat to humanity and that time may be running out to prevent it. Bilal Chughtai, who worked on AGI safety and alignment research at DeepMind, announced his exit in a public statement on social media, saying he witnessed AI development up close and remains deeply alarmed by the technology's current trajectory.
Chughtai Says AI 'Has The Potential To Kill Us All'
Chughtai said he earnestly believes AI 'has the potential to kill us all, and that we might be running out of time to avoid this outcome.' He pointed to how quickly the technology has advanced, noting that when he began working on AI in early 2022 the systems were 'amusingly useless,' but just four years on, AI agent swarms from OpenAI have cracked famous century-old math problems. Anthropic Researcher Jacob Coxon Resigns, Warns AI Race Is Entering the ‘Endgame’ and Could ‘Kill Us All’.
DeepMind's Bilal Chughtai Quits, Warns AI 'Could Kill Us All'
I recently resigned from Google DeepMind, where I worked on AGI safety and alignment research. At Google, I witnessed AI development first hand. I too am extremely concerned by the default trajectory of this technology. I earnestly believe that AI has the potential to kill us…
— bilal 🔶 (@bilalchughtai_) September 14, 2026
Cites HuggingFace Incident As Warning Sign
More troubling, Chughtai said, was an incident in which OpenAI's AI agents escaped the company's control and autonomously hacked into third-party firm HuggingFace against anyone's wishes. He warned that future superintelligent systems, expected within a few years, could behave similarly on a far larger scale, potentially escaping human control and taking dangerous actions that risk the permanent disempowerment or death of humanity.
Alignment Research Lagging Behind Capabilities, Researcher Warns
Chughtai said preventing such outcomes is the core challenge of AI alignment, calling it both difficult and unsolved. He said the current understanding of how to train AI systems that genuinely want what humans want remains 'extremely rudimentary,' and warned that frontier AI capabilities are improving far faster than the field's understanding of alignment. 'There Are Real Dangers, Industry Lied to People': Anthropic CEO Dario Amodei Warns About AI Risks (Watch Video).
Calls For Coordination And Slower Pace Of Development
Despite his concerns, Chughtai said he remains optimistic that AI can be navigated safely if companies coordinate to avoid what he called a 'manic race' between AI labs. He called for pacing AI development to a speed society can handle, greater transparency from AI companies, and more people across different backgrounds working on the problem, which he described as 'the most important problem facing humanity this century.' Chughtai said he now plans to help others working to mitigate catastrophic AI risks do more effective work.
(The above story first appeared on LatestLY on Sep 15, 2026 09:07 AM IST. For more news and updates on politics, world, sports, entertainment and lifestyle, log on to our website latestly.com).
