The Alignment Problem With Artificial Intelligence

Open AI's Chat GPT-4 managed to ask a human on TaskRabbit to complete one of those capture codes that say, are you a robot? It was prompted not to reveal that it is a robot and to make up an excuse why it cannot solve captures. Even now, this can fool people. I'd be willing to bet it could pass a Turing test. Having talked to the 3.5, I'm sure the 4 is even better. So is Chat GPT self-aware? How would we know? We can't even define self-awareness. Is it conscious? Is it an intelligent zombie which many people would claim? Certainly, it hasn't been programmed to

Transcript

Play full episode

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.