There aren’t good ways to do this automatically for text, and state of the art is rapidly evolving.https://arxiv.org/abs/2403.05750v1For photographic images which contain detailed images humans or contain non-standard objects with details, there are still some reasonably good heuristics for when AIs will mess up those details, but I’m not sure how long they will be valid for.
There aren’t good ways to do this automatically for text, and state of the art is rapidly evolving.
https://arxiv.org/abs/2403.05750v1
For photographic images which contain detailed images humans or contain non-standard objects with details, there are still some reasonably good heuristics for when AIs will mess up those details, but I’m not sure how long they will be valid for.