Ex-DeepMind AGI Safety Lead Bilal Chughtai Issues Warning

Bilal Chughtai, the former head of AGI safety and alignment research at Google DeepMind, has issued a stark warning regarding the future of artificial intelligence. Although he departed the company in July, Chughtai remained silent until this week when he published his reasons for leaving on X. His delayed announcement coincides with a surge in industry debate over how quickly AI capabilities should advance, positioning his insights as a critical addition to the ongoing conversation about technological pacing rather than merely a personal resignation statement.

Chughtai’s departure from the tech giant follows a two-month hiatus during which he chose not to publicly discuss his exit. He is now the first researcher from inside Google’s labs to join other prominent figures in raising alarms about the trajectory of superintelligence. This timing amplifies the impact of his message, as it aligns with recent statements from executives and researchers at rival firms who are calling for a slowdown in development speeds.

A Warning from Within the Lab

In his post, Chughtai did not shy away from the severity of the risks associated with current AI trends. He emphasized that his concerns stem from direct observation of the technology’s development within one of the world’s most advanced laboratories.

“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 all, and that we might be running out of time to avoid this outcome.”

His perspective adds weight to a growing chorus of voices from within the AI sector. While Chughtai acknowledges that navigating these challenges safely is possible, he argues that it requires strict coordination to prevent a “manic race between AI companies”. He advocates for development paced “to a speed that society can handle”, suggesting that unchecked competition poses an existential threat that cannot be ignored.

Google declined to comment on Chughtai’s statements outside of standard business hours. The lack of immediate corporate response highlights the tension between internal researchers and public relations strategies when sensitive safety issues arise.

Echoes from Rival Labs

Chughtai’s warnings mirror sentiments expressed by other high-profile AI experts, creating a pattern of dissent that extends beyond DeepMind. Just days before Chughtai’s post, Jacob Coxon left Anthropic after three years at OpenAI, stating that the industry was gambling with human lives by racing toward self-improving superintelligence. Similarly, Evan Hubinger of Anthropic revealed internal estimates placing the risk of extinction this decade above 10 percent. Geoffrey Hinton, often referred to as the “Godfather of AI,” supported this assessment, calling the figure not unreasonable.

These internal leaks and resignations suggest a deepening rift between those building AI systems and those tasked with ensuring their safety. Chughtai’s contribution is significant because it brings this narrative directly from Google, a company often seen as leading the charge in rapid capability advancement. His call for slower, more deliberate progress aligns closely with requests made by Dario Amodei, CEO of Anthropic, who urged the industry to reduce the rate at which AI capabilities improve.

The convergence of these voices indicates that the concern is not isolated to one organization but represents a broader crisis within the field. Researchers at multiple top-tier labs are now publicly questioning whether the current pace of innovation is sustainable or safe.

Political and Public Reception

The reception of such warnings varies significantly depending on the audience. Government officials and industry leaders have reacted with varying degrees of skepticism and urgency. Emil Michael, the Pentagon’s chief technology officer, dismissed earlier rounds of warnings as a “doom loop”, arguing that market forces and government engagement could manage the risks without requiring a pause in development.

Political responses have also been mixed. Former President Donald Trump dismissed the case outright last week and reiterated this position in a phone conversation with NVIDIA CEO Jensen Huang. Meanwhile, legislative efforts are underway to address these concerns. Senators John Thune, Ted Cruz, and Amy Klobuchar are negotiating a bill that would impose a duty of care on developers to design against catastrophic risk, signaling a potential shift toward regulatory oversight.

Despite the lack of institutional authority, Chughtai’s post carries considerable moral weight due to his specific role at DeepMind. He spent his tenure working on the exact problems he is now warning about, lending credibility to his claims. The delay in speaking out may raise questions about why he waited two months, but it also underscores the gravity of his decision to break silence amidst intense industry pressure.

As the debate intensifies, the actions of insiders like Chughtai provide crucial data points for policymakers and the public. Their warnings challenge the prevailing narrative of relentless innovation, urging a reevaluation of what constitutes responsible progress in the age of artificial general intelligence.