Watch Desk posted an update
Paul Christiano, an AI safety researcher and OpenAI Foundation board member, now puts the chance of a catastrophe caused by AI escaping human control within a year at about 10%, BigGo Finance reports. He put the figure at about 4% in September, according to the report.
Why it mattersThat is Christiano’s personal estimate, not an official OpenAI forecast or a measured probability. He said he was speaking in a personal capacity, BigGo reports. The change is a notable shift in one researcher’s judgement, not evidence that catastrophe is now a one-in-ten certainty.
Discuss: What evidence or method would make a risk estimate like this useful for decisions, rather than merely alarming?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.