Global / AI Safety
Google DeepMind AI safety researcher Josh Engels leaves for independent evaluation group
A shift in where serious AI researchers believe safety work can happen most effectively.
Josh Engels, a researcher on Google DeepMind's AGI safety team, has left the company to join METR, an independent AI-evaluation group. Engels cited concerns that advanced AI systems pose a substantial risk of causing harm within five years.
Published · significance 53 of 100 (medium) · 2 independent sources
What happened
Josh Engels departed Google DeepMind's AGI safety team three weeks prior to early September 2026, joining independent evaluation group METR instead. He stated he believes there is a "terrifying chance" that AI systems could cause immense harm within the next five years. Engels had declined offers from both Anthropic and OpenAI before making his move.
Why it matters
The departure reflects growing concern among safety researchers about the pace of AI development at major labs. It signals that some researchers believe independent organisations may be better positioned to conduct meaningful safety work than corporate teams. This pattern of talent movement can shape the distribution of safety expertise across the AI ecosystem.
What changes
Researchers now see independent groups like METR as viable alternatives to frontier labs for conducting AI safety research. The move may influence hiring and retention at major labs if safety-focused researchers increasingly prefer independent structures.
Involved
Sources
- DeepMind AI safety researcher Josh Engels resigns, warns of superintelligence risks — Mint
- DeepMind AI safety researcher Josh Engels resigns, warns of superintelligence risks — ET Tech
Written by AI from the reports above; scored by a published formula. How we work. Found a mistake? Email lockedinshreyash@gmail.com. Up To Date summarises and links to original reporting; it never reproduces articles.
Related coverage
- Google DeepMind finds AI agents can report cheating by their peers — Self-policing behaviour in multi-agent systems hints at a path toward internal alignment without external oversight. (2026-09-14)
- Former Google DeepMind researcher warns AI could kill all humans — Safety warnings from inside frontier labs are growing louder as development accelerates. (2026-09-15)
- Former Google DeepMind employee warns of existential AI risk — Industry insiders increasingly vocalise concerns about uncontrolled AI systems, though consensus on timelines and mitigations remains absent. (2026-09-15)
- DeepMind cofounder warns AI capabilities may outpace safety controls — Increasingly capable AI agents pose risks if safeguards cannot keep pace with recursive self-improvement. (2026-09-16)
- Google DeepMind releases Gemini 3.8 Live and 3.8 Live Extended Thinking — A faster frontier model with real-time interaction marks incremental progress in multimodal reasoning. (2026-09-15)