Science at LLM Speed

"Sometimes the papers are real. Sometimes a real paper is attached to a claim it does not support. Sometimes the title, authors, or venue are slightly wrong. And sometimes the entire reference was generated because a plausible citation completed the paragraph."

This is something I've been also saying many times; it is not sufficient to check whether the meta-data (bibliography) is correct. And the meta-data should nowadays come straight from Crossref anyway.