AI researchers warn that superintelligence could threaten human survival
A group of AI researchers has published video interviews warning that artificial superintelligence could pose a severe danger to humanity. Geoffrey Irving, a former employee of OpenAI and Google DeepMind, said he viewed the possibility of human extinction as roughly a coin flip.
The interview is one of about a dozen conversations with current and former researchers from OpenAI, Google and Anthropic. Several participants joined Irving in raising the possibility that advanced AI systems could cause humans to become extinct. Neel Nanda, a research scientist at Google DeepMind, said there was “at least a 10 percent chance” that AI could cause human extinction.
Palisade Research, which describes itself as a nonprofit focused on studying AI capabilities and motivations, organized the interviews and released them through frominside.ai. The project presents researchers’ assessments of the potential risks associated with superintelligence, rather than a single shared forecast.
Written by Sakhanda Wire AI from reporting by The Verge. Facts, quotes and full context are in the original article below.