AI researchers put out videos saying superintelligence is ‘exactly as hazardous as it sounds’

“The chance of individual extinction is concerning a coin flip, in my view,” Geoffrey Irving, a erstwhile OpenAI and Google DeepMind employee, stated in a new interview. It’s one of a dozen interviews alongside AI researchers, including current and erstwhile workforce at OpenAI, Google, and Anthropic. Palisade Research, which says it’s a nonprofit studying AIs’ capabilities and motivations, collected and launched the interviews on frominside.ai.

Several of the researchers joined Irving in alert concerning the possible AI could run humans extinct. Neel Nanda, a Google DeepMind investigation scientist, stated there’s “at smallest a 10 percent chance that it causes individual extinction, and that is ridiculously high.” We noted comments from the two Nanda and Irving in a latest study concerning the “AI safety” community, and how difficult it is to nail downward what that term means, or what anyone should do concerning it, and equal these videos don’t have one cohesive resolution to offer.

Former OpenAI investigator Daniel Kokotajlo went as far as claiming that superintelligent AI “would basically be god-like powerful,” and added, “unfortunately we don’t cognize how to authority them at all. And so likely nobody would authority them. This is exactly as hazardous as it appears and must not be allowed to happen.”

While all video is accessible in full, the location additionally collects segments from multiple group speaking on akin topics, including whether this is all fair hype or marketing, and why they would activity on item that they accept is this dangerous. Some of the responses contain Google’s Mary Phuong saying “I think you entirely have to be doubtful of what I’m saying since I am being paid by the lab,” and Nanda’s clarification that “If I did not accept that my activity was immediately reducing these existential risks, I would quit… These companies are not going to halt making these systems fair since I quit.”