4 ms·

These sort of agent-architecture AIs are where I think things will head and also where things will get more dangerous. When an AI has a goal and an ability to

by zoba 4y ago

These sort of agent-architecture AIs are where I think things will head and also where things will get more dangerous. When an AI has a goal and an ability to break the goal down, and make progress towards the goal… the only thing stopping the AI from misalignment is whatever it’s creator has specified as the goal. Things get even more tricky when the agent can take in new information and deprioritize goals. I have been curious why OpenAI haven’t discussed Agent-based AIs.

jazzyjackson 4y ago
> the only thing stopping the AI from misalignment is whatever it’s creator has specified as the goal. And the API providers willingness/capability to provide the service being requested. I'm with those who don't see the danger, besides humans' own stupidity being amplified and someone trusting "AI" with something life critical, what should I be watching out for?