No.
> How do we make sure that AI does not fake them by silently (by silent changes or by faking the test output)?
I have never seen this happen for a couple of months at least.
They're representative of scenarios which your application can enter.
So if the tests were mocked and a bug went live he'd still notice, after all it means your software behaved incorrectly and now you gotta clean up the pieces.
The discussion went from tests being written by ai, to nobody actually reviewing the tests, to the claim that given those circumstances, you wouldn't notice that the tests may not tests anything anymore because of mocks.
My statement was merely pointing out that your previous comment was groundless because they still would've found out about those bugs when they went live.
Maybe. You're assuming visible and obvious bugs.