We do have code reviews for pull requests. But on average I would guess there is great amount of complacency there. I suppose old proper QA phase was best answer we had. But that is expensive and slow.
i mean i don't disagree with you, but then there's at least unit test cases that we write as devs in a certain way that ai just doesn't understand quite well yet. we worked so much on speed that we forgot along the way that slow was a safety net. we tend to let our qa teams leverage ai as much as they want to automate test cases but i don't feel super comfortable getting rid of it just yet