AI researchers put out videos saying superintelligence is ‘exactly as dangerous as it sounds’
A dozen current and former AI researchers from OpenAI, Google, and Anthropic are now publicly stating existential risk estimates — and some put human extinction odds at a coin flip.
Why it matters
AI safety researchers embedded in frontier labs are breaking ranks on existential risk in coordinated video interviews. This signals internal concern translated into public testimony, but the estimates remain ungrounded in measurable capability milestones or deployment failure modes — practitioners should note the gap between researcher conviction and operational risk frameworks.
The key facts
11 to knowGeoffrey Irving (ex-OpenAI, ex-Google DeepMind): ~50% human extinction probability in interview
Neel Nanda (Google DeepMind research scientist): at least 10% extinction probability
12 interviews collected by Palisade Research nonprofit, published on frominside.ai
Current and former employees from OpenAI, Google, Anthropic participate
No capability benchmarks, deployment incidents, or technical mechanism cited to ground probability estimates
Published September 29, 2026
Geoffrey Irving (ex-OpenAI, Google DeepMind): ~50% chance of human extinction
Neel Nanda (Google DeepMind): ≥10% chance of human extinction
Interviews collected by Palisade Research, published on frominside.ai
Multiple current and former employees at OpenAI, Google, Anthropic participating
Framed as researcher consensus on danger; no new technical or safety discovery cited
Go to the source
The Verge AItheverge.com
Publisher excerpt: "The chance of human extinction is about a coin flip, in my view," Geoffrey Irving, a former OpenAI and Google DeepMind employee, said in a new interview. It's one of a dozen interviews with AI researchers, including current and former employees at OpenAI, Google, and Anthropic. Palisade Research,…