“In my opinion, the possibility of human extinction is a coin flip,” 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, which claims to be a nonprofit that studies the capabilities and motivations of AIs, compiled and posted the interviews on frominside.ai.
Several of the researchers joined Irving in warning about the possibility that AI could make humans extinct. Neel Nanda, a research scientist at Google DeepMind, said there is “at least a 10 percent chance it will cause human extinction, and that’s ridiculously high.” We noted Nanda and Irving’s comments in a recent report on the “AI safety” community, and how difficult it is to pin down what that term means, or what anyone should do about it, and even these videos don’t have a coherent solution to offer.
Former OpenAI researcher Daniel Kokotajlo went so far as to claim that superintelligent AI would “basically be powerful as a god,” adding, “unfortunately we don’t know how to control them at all. Therefore, probably no one would control them. This is exactly as dangerous as it sounds and should not be allowed to happen.”
While each video is available in its entirety, the site also compiles segments from various people talking about similar topics, including whether this is all just advertising or marketing, and why they would work on something they believe is so dangerous. Some of the responses include Google’s Mary Phuong saying, “I think you should absolutely be suspicious of what I’m saying because the lab pays me,” and Nanda’s explanation that “if I didn’t believe my job was directly reducing these existential risks, I would quit… These companies aren’t going to stop making these systems just because I quit.”
