We have to stop ignoring AI’s hallucination problem

misk@sopuli.xyz to Technology@lemmy.world – 496 points –
We have to stop ignoring AI’s hallucination problem
theverge.com
203

You are viewing a single comment

The Chinese Room thought experiment is a good place to start the conversation. AI isn't intelligent, and it doesn't hallucinate. Its not sentient. It's just a computer program.

People need to stop using personifying language for this stuff.

that's not fun and dramatic and clickbaity though

I think it's that, and something even worse as well. There are probably many well meaning people working on these things thinking they really are creating and guiding and intelligence. It's an opportunity to feel like a god and a tech wizard at the same time.

Technically, humans are just bio machines, running very complicated software. AI just isn't there yet.