Post

In spot tests I've done, I see this happen *all of the time*, such as with my own papers. The LLM would often rather hallucinate a male scientist than get my name right. 🤷‍♀️ It's kind of unsurprising that models trained unscrupulously on the entire internet are misogynistic

Mel Andrews@bayesianboy.bsky.social

When utilized in literature review, LLMs consistently 1. fail to mention female authors in female-led literatures, 2. insist that men are more influential or more heavily cited when this is contradicted by objective citation counts, and 3. attribute women’s work to hallucinated male scholars.

17/6/26, 6:07 am205 reposts583 likes

Dr. Cindy Conaway

@drcindyconaway.bsky.social

17 June

It cited an article I co-wrote as by Harold Conaway. Who’s that?

ecleclecl.bsky.social

@ecleclecl.bsky.social

17 June

I have a screenshot of Google's AI Overview doing this to me! Emilio Colunga...😒

Emily Hunt

@emily.space

17 June

receipts ⤵️

Emily Hunt@emily.space

thought I'd give it another chance. now i'm an andrew lol (and I have never written a paper about Kepler before so it's wildly wrong)

Emily Hunt

@emily.space

17 June

Another source ⤵️

It's ME(Jaime)@exceedhergrasp1.bsky.social

A source on this: arxiv.org/abs/2508.02740

arxiv.orgWho Gets Cited? Gender- and Majority-Bias in LLM-Driven Reference SelectionLarge language models (LLMs) are rapidly being adopted as research assistants, particularly for literature review and reference recommendation, yet little is known about whether they introduce demogra...

atomichfish.bsky.social

@atomichfish.bsky.social

23 July

If the mathematical data regurgitator machine is sexist, isnt this more of the problem with the data its regurgitating rather than the algorithm itself?

Matthew 🦖

@moosewithablanket.bsky.social

17 June

The most insufferable people I've encountered regarding LLMs act like the bigotry is just objective reality. AI companies know that they can't get rid of this behavior. They can only partly mitigate it. And it's not always going to be easy to tell when it has affected the output.