Search
On July 17, Demis Hassabis, CEO of Google DeepMind, proposed the establishment of an international independent body to conduct unified testing and review of advanced AI models before their release. He plans to visit Washington next week to discuss the proposal with the U.S. government. Hassabis highlighted that the advanced cyberattack capabilities demonstrated by Anthropic's Mythos model underscore the urgency of strengthening AI safety oversight. The proposed plan would involve independent experts conducting reviews, with leading AI companies providing financial support. The testing period for models would be up to 30 days, with coordination across the industry to slow down the development of advanced models if necessary. Currently, OpenAI CEO Sam Altman and Elon Musk have publicly expressed support for the proposal.
-
C114 Communication Network -
Communication Home
