Post

What's most upsetting me as a journal editor is that many papers we receive where AI use is disclosed for editing and language refinement actually appear to have been ruined rather than improved by AI. We get what seem like promising underpinning studies, but totally obscured by over-complex text.

116 reposts518 likes

This is different to the lazy auto-generated synthetic manuscripts we frequently receive. The ones I mean often sadly look like PhD or early career projects where the data seem good but chat or claude has garbled the presentation of the analysis. Writing reject letters for these is hard work.

I wish our sector could support academic writing better, especially for early career and non-anglophone writers. Otherwise they end up punished with rejection when they turn to AI, convinced by current discourse that it's the solution to academic communication. Seen too many papers ruined by AI now.

As an editor, I’ve been trying to tell people that AI text replaces easy-to-fix problems with hard-to-fix ones. But my employer is all in on AI, so it’s hard to get the message out.

This is indeed a problem. But my experience is even worse: getting a dozen or more papers each week submitted by "independent researchers" with weak results showing little or no understanding of what makes a publishable mathematics article.

On the other hand, in the hands of an experienced mathematician who understands the literature and knows what problems are interesting and amenable, LLM's are now being used routinely to solve good, hard problems that have been open for years or even decades. It's a revolution.

It’s a whole new world for editors and teachers. In “Memories, Dreams, and Reflections,” Carl Jung tells of an incident where he was accused of plagiarism because his essay was of such high quality that the teacher was certain it was copied but could not find the source. It nearly destroyed him.

Absolutely agree. Reviewed an article recently where authors declared AI assistance for refining language and for image generation, they claimed to have checked everything and took full ownership. The number of spelling mistakes in the figures alone called into question validity of the whole paper

AI is an excuse to not think and refine your thoughts. My posts are rife with bad typing, logical fallacies, alcohol abuse in the wee hours, sheer and raw stupidity, callousness, bad taste, but I would never degrade my words by abdicating my "wit" to AI. For better or worse, this is me. I am real.

That matters. We must learn to reason on our own, and we must grow into who we become instead of accepting an identity of amalgamated human paste, derived by processed thought, devoid of caloric or nutritional value. Be real. Be flawed. Grow from it.