AI models are using material from retracted scientific papers

Who would've thought (me)

Yeah, no shit? LLMs don't actually know or understand anything, the idea that a scientific paper was retracted means nothing to them, all the training cares about is the patterns of word usage. It doesn't matter that some core part of the paper was wrong, it was SAID, and it is presumed to be grammatically correct.

This is a core reason why using LLMs as a search engine is fucking stupid, they don't filter for what's accurate, only what has been said before.

The future is hydroxychloroquine and Tylenol.

Once again, the statistical word parrot does not have any understanding of what it's doing or what it's saying or have any ability to "learn" new things.

Retracted papers? They cite papers that never existed

To be fair, human scientists still use material from retracted scientific papers too