OpenAI staff member roon notes Asimov did not consider adding more laws for robots
OpenAI technical staff member roon posted on X that Isaac Asimov had not considered adding more laws when developing rules for robotic behavior and its use in modern AI systems. AI safety researcher davidad quote-tweeted the post and linked to his own update. That update stressed Asimov's preference for narrative flexibility over strict robustness and included a revised set of laws for practical robotic systems.
#458davidad 🎇@DAVIDAD@tszzl Asimov wasn’t even trying
#458davidad 🎇@DAVIDAD@tszzl
@TheZvi @tszzl That’s a good point. Asimov *did* try adding another law that time!
I actually haven’t read it, but the resulting robot, Daneel Olivaw, has come up in my LLM-assisted exploration of Anglophone literary precedent for artificial bodhisattvas, as probably the single closest match.
Zvi Mowshowitz@TheZvi@davidad @tszzl I think he was trying in Robots and Empire and... well, that's one answer.
@davidad @tszzl I think he was trying in Robots and Empire and... well, that's one answer.
davidad 🎇@davidad@tszzl
@davidad @tszzl It's arguably the only book in the entire Asimov universe that actually matters...
davidad 🎇@davidad@TheZvi @tszzl That’s a good point. Asimov *did* try adding another law that time! I actually haven’t read it, but the resulting robot, Daneel Olivaw, has come up in my LLM-assisted exploration of Anglophone literary precedent for artificial bodhisattvas, as probably the single closest match.