Comment by nazgul17
1 day ago
The thing is, when you copy paste a bibliography entry from the publisher or from Google Scholar, the authors won't be wrong. In this case, it is. If I were to write a paper with AI, I would at least manage the bibliography by hand, conscious of hallucinations. The fact that the hallucination is in the bibliography is a pretty strong indicator that the paper was written entirely with AI.
Google Scholar provides imperfect citations - very often wrong article type (eg article versus conference paper), but up to and including missing authors, in my experience.
I've had the same experience. Also papers will often have multiple entries in Google Scholar, with small differences between them (enough that Scholar didn't merge them into one entry).
I'm not sure I agree... while I don't ever see myself writing papers with AI, I hate wrangling a bibtex bibliography.
I wouldn't trust today's GPT-5-with-web-search to do turn a bullet point list of papers into proper citations without checking myself, but maybe I will trust GPT-X-plus-agent to do this.
Reference managers have existed for decades now and they work deterministically. I paid for one when writing my doctoral thesis because it would have been horrific to do by hand. Any of the major tools like Zotero or Mendeley (I used Papers) will export a bibtex file for you, and they will accept a RIS or similar format that most journals export.
This seems solvable today if you treat it as an architecture problem rather than relying on the model's weights. I'm using LangGraph to force function calls to Crossref or OpenAlex for a similar workflow. As long as you keep the flow rigid and only use the LLM for orchestration and formatting, the hallucinations pretty much disappear.