Elektrine lite

← Feed

@foaylward@genomic.social

Post #937042

2026-03-29 19:24 UTC

Some colleagues have told me that reviews they received on manuscripts were likely AI-generated, so for journal club last week we asked Claude to write both a critical and a positive review of the paper we discussed. It was written by a former postdoc in the lab, so definitely a field familiar to us. Both of the reviews we got were essentially fancy-sounding BS with no substance. It was simply a regurgitation of material discussed in the manuscript already. #ai #academia #science

Replies (1)

  • @foaylward@genomic.social 2026-03-29 19:26

    Important comments that a good human reviewer would have provided, such as recommending alternative software, an alternative interpretation, etc, were completely absent. It struck me as pretty easy to identify reviews made this way. But if it wasn't an area that I was less familiar with, it may have required a bit more time to spot it.

    Open ##2832149