[jammies]

To me, it doesn't really matter if the model is infallible, it just matters if the model is as likely to be correct as an expert human. It should be used like any other source of information, you need to cross check against other sources before taking its conclusions as truth. But that is still extremely useful, considering expertise is expensive and this is comparatively cheap, at least to the end user.

Of course people are going to misuse it and trust it for things that it shouldn't be trusted in. That's because people are often flawed, ignorant, and lacking rigorous thought processes. Unlike these narrow AI models, though, they are also unlikely to collectively improve.

Sent from my Pixel 4a (5G) using Tapatalk