Palisade Research, a nonprofit studying AI capabilities and motivations, has published a dozen interviews with current and former researchers from OpenAI, Google and Anthropic. Several warn that advanced AI could pose an existential threat, bringing renewed attention to disagreements over how serious the risk is and what, if anything, should be done.

The researchers offer sharply different estimates and explanations: former OpenAI and Google DeepMind employee Geoffrey Irving puts the chance of human extinction at roughly 50 per cent, while Google DeepMind’s Neel Nanda estimates at least 10 per cent. Former OpenAI researcher Daniel Kokotajlo says superintelligent AI could be extremely powerful and difficult to control; the videos also address whether such warnings are hype and why researchers remain in the field.