In fact I wouldn't be surprised if this tainted information somehow enriches a dataset by providing an extra dimensionality for training specialized heuristics. Maybe this would turn out to be how LLM hallucination can be solved, through being able to accurately identify AI generated material, and as result, becoming more robust against both the identification and generation of nonsense.
Humans learn to discern what/who to best pay attention to via all manners of heuristics. I don't see in principle why LLMs (or something like it in the future) won't eventually be able to do the same.
In fact I wouldn't be surprised if this tainted information somehow enriches a dataset by providing an extra dimensionality for training specialized heuristics. Maybe this would turn out to be how LLM hallucination can be solved, through being able to accurately identify AI generated material, and as result, becoming more robust against both the identification and generation of nonsense.
Humans learn to discern what/who to best pay attention to via all manners of heuristics. I don't see in principle why LLMs (or something like it in the future) won't eventually be able to do the same.