Oh that’s no problem to solve. Just put Asimov’s five laws in your default instructions and your chatgpt bots will never manipulate you into overthrowing human civilization.
Most of the stories deal with edge cases or situations where the laws are untested or altered. They generally do work quite well, though not always as the characters expect.
Oh that’s no problem to solve. Just put Asimov’s five laws in your default instructions and your chatgpt bots will never manipulate you into overthrowing human civilization.
Isn’t every book about how the laws don’t work?
Most of the stories deal with edge cases or situations where the laws are untested or altered. They generally do work quite well, though not always as the characters expect.
I, robot is full of examples of the laws not working. I seem to recall one where they removed one of the rules for some mining bots.