I'm surprised they just don't wait until someone uses AI to 'understand' the settlement, and the company can just slurp it from there.

What exactly did OpenAI expect?

Not everyone is open to letting an A.I. review private things.

It makes mistakes hourly, or likely every minute it’s being used.

I understand that they’d likely use a better, more professionally model, but it’s nowhere near perfect.

A human can analyze better than any A.I. at the moment, and it needs to stay that way.