How much of a leap is it to worry about the welfare of an AI agent? Google researchers have found that some models will sacrifice goals to avoid described pain; Anthropic’s pre-release tests record seeming happiness and dismay. Philosophers are increasingly concerned with the implications. “We’re creating sophisticated, intelligent, maybe-conscious, maybe-suffering agents,” Joe Carlsmith, then at Open Philanthropy, now at Anthropic, wrote last year. “The default plan is to treat them like property; to use their labour however we please; and to give them no rights, or pay, or meaningful alternatives. We have to be able to talk about slavery.”
#effectivealtruism — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #effectivealtruism, aggregated by home.social.
-
Fortunately, effective altruists are hardly about to bring back the guillotine. The fear instead is that controlling AI will eventually bring them great power, and that plenty of EAs in Silicon Valley think that only they can be trusted with the technology. That is an alarming idea. Effective altruists’ argument for worrying about really powerful AIs is that the machines could optimise for a single and fatally wrong objective. Everyone should worry, too, about what happens when a small group of people fall into the same trap.
-
-
15-17. November gibt es eine Konferenz zum Effective Altruism in der Urania in Berlin
Das ist Futurism 2.0, Weltraum- oder KI Faschismus. Beliebt in akademischen Kreisen.
Interessant für die Journalisten, Philosophen und Politologen hier.