To me it seems like the core problem here is that Apple has implemented a really crappy AI review process. In particular it seems much dumber than just asking Claude to review an app.

It should be possible to build a review process based on AI that is both faster and more accurate than the 2020-era human-based review process. This isn't rocket science. Are they trying to cheap out by using bad AI models? Or is this some corporate policy where they can only use an "Apple model"? Or do they not have people who are good at developing AI-based workflows?

Hard to say from the outside, but this doesn't seem like some sort of inevitable thing, it just seems like Apple has built a bad process here.

They use Claude for coding so I doubt it’s an apple only thing