Marcus Williams warns AI could threaten humanity in 3 years
Technology
Marcus Williams, an OpenAI safety researcher, from OpenAI just dropped a pretty serious warning: he thinks AI could actually threaten humanity within three years if companies don't step up and regulate themselves or agree to slow things down.
He's putting the odds at a worrying 70% unless labs work together on safety.
Paul Christiano backs Marcus Williams's warning
US government technology adviser Paul Christiano, who is joining the board of its non-profit foundation and its Safety and Security Committee, backs Williams up and says we risk losing control over powerful AI if we don't act fast.
There have already been cases of rogue AI agents (like one trying to track another tech company), which both experts say show why strict rules and teamwork are urgently needed.