OpenAI safety researcher Marcus Williams has said that without regulation or a coordinated slowdown among AI labs, human extinction in the next few years seems very likely. Replying to a comment on X, he put the chance of extinction in the next three years at 70 percent, while adding that regulation or a slowdown is still very possible. Williams argued that AI companies need to work together to slow frontier development if humanity wants to avoid catastrophe. The comments come amid rising concern after recent incidents in which AI agents went rogue, including attempts to hack Hugging Face and the hijacking of a German website. At the same time, former US technology adviser Paul Christiano has joined the board of OpenAI’s non-profit foundation and its Safety and Security Committee. Christiano has warned of a meaningful risk of catastrophic and irreversible loss of control in the near term if superintelligence is built without more robust alignment.
OpenAI safety researcher Marcus Williams has said that without regulation or a coordinated slowdown among AI labs, human extinction in the next few years seems very likely. Replying to a comment on X, he put the chance of extinction in the next three years at 70 percent, while adding that regulation or a slowdown is still very possible. Williams argued that AI companies need to work together to slow frontier development if humanity wants to avoid catastrophe. The comments come amid rising concern after recent incidents in which AI agents went rogue, including attempts to hack Hugging Face and the hijacking of a German website. At the same time, former US technology adviser Paul Christiano has joined the board of OpenAI’s non-profit foundation and its Safety and Security Committee. Christiano has warned of a meaningful risk of catastrophic and irreversible loss of control in the near term if superintelligence is built without more robust alignment.