Palisade Research posts AI extinction warnings

Palisade Research, which describes itself as a nonprofit studying AIs' capabilities and motivations, gathered the interviews and launched them on frominside.ai. The set includes current and former employees of OpenAI, Google, and Anthropic, so the warnings come from people with direct ties to the labs building the systems in question. The site also collects segments from multiple speakers on similar topics, including whether this is all just hype or marketing.

Several of those segments address the obvious objection: why would someone work on something they describe as this dangerous? Google's Mary Phuong says, 'I think you absolutely should be suspicious of what I'm saying because I am being paid by the lab.' Neel Nanda's answer is that 'if I did not believe that my work was directly reducing these existential risks, I would quit… These companies are not going to stop making these systems just because I quit.' Those two answers point in different directions — one invites skepticism, the other justifies staying.

What the interviews notably do not offer is a single cohesive solution, even among people who agree the risk is real. That gap is likely what the project is meant to expose: the disagreement is not only about whether AI is dangerous, but about what anyone should do about it. Whether the videos change how labs or their staff handle these warnings may depend less on the numbers cited than on whether the people saying them are seen as representative or as outliers.