Hallucinated Citations and the Matthew Effect

A month or so ago, I heard about a study that found 85% of LLM-hallucinated citations in scientific article preprints were also in the later journal versions. That means they were not caught by peer review. Bad, right?

It gets worse. The fake citations also disproportionately gave credit to male scholars, who already have many citations. In the BlueSky post where I saw the study mentioned, this was called "a fake-citation Matthew effect."

I had never heard of the Matthew effect, so looked it up. The term goes back to 1968. Robert Merton and Harriet Zuckerman named it (ironically, Merton is the one usually credited with it). 

The name comes from the Parable of the Talents from the Gospel of Matthew in the Bible. In general, it's the tendency for the rich to get richer, but in academia, it's the way "eminent scientists will often get more credit than a comparatively unknown researcher, even if their work is similar; it also means that credit will usually be given to researchers who are already famous."

So the fact that LLM-generated fake citations credit more-established, male scholars is just one more thing to hate about them. They can't even hallucinate fairly.