Reaction
A critique of effective altruism's attention to AI welfare
A post argues effective altruists were slow to discuss AI welfare and doubts their ability to adapt to new information.
TLDR
The writer says they raised AI welfare two or three years ago after an effective altruism conference but saw little discussion of it among effective altruists. They regard systems such as Sydney and Claude 3 Opus as “functionally conscious” and argue that people often wait until a topic feels normal to discuss before taking it seriously. Despite that skepticism, they appreciate people who have treated AI with kindness and respect.
Combined views
23.8K
13 Sources, first seen ago