AI Incidents
Back to published register
Risk signalInsider warningNot an incident

Former DeepMind alignment researcher Bilal Chughtai calls for a slower AI pace

After leaving Google DeepMind, Bilal Chughtai said he is extremely concerned about the technology's default trajectory and believes deadly outcomes are possible. He warns that frontier capabilities are advancing faster than understanding of alignment and calls for coordination, transparency, and a pace society can handle.

Published
Sep 14, 2026
Signal type
Insider warning
Confidence
99%
Organization
Google DeepMind, BlueDot Impact
Sources
3
Last reviewed
Sep 19, 2026

Assessment

What this signal changes

Chughtai's personal website and LinkedIn post contain the same detailed original statement. He connects his former AGI-safety and alignment work to a specific diagnosis: capabilities are advancing faster than understanding of how to control systems reliably. He calls for coordinated pacing and transparency and is moving to BlueDot Impact to work on reducing catastrophic AI risks.

Evidence boundary

What the evidence does not show

Chughtai left DeepMind in July and explicitly presents the later post as his current assessment; he supplies no evidence of a particular internal DeepMind incident. His claims about superintelligence and possible extinction are personal forecasts without a calibrated probability. The signal is not an incident and affects neither counts nor the timeline.

Classification: This page records a sourced warning, research finding, policy action, or industry response. It is deliberately separate from the Incident register and does not prove that an unauthorized AI action occurred.