Arxiv, the popular preprint server, recently decided to penalize submitters of articles that contain non-existing references, presumably hallucinated by an AI. All authors will be sent to the doghouse, not just the first author. They’ll be blocked for a year from submitting new articles and will have their whole history reviewed, among other sanctions.
While this new rule was generally met with support, some hapless US professor decided to publicly disagree and complain that nobody could expect him to be held responsible for the whole paper, including fake references that other authors may have added.
The to-be-expected shitstorm ensued.
You may wonder what the professor was thinking. I don’t know, but it is obvious to me where he is coming from. It is a larger problem of academic fraud than just hallucinated references. Pretty much every university has this problem.
Some professors are predominantly science managers rather than traditional professors. In particular in Germany, in engineering, there are professors publishing hundreds of papers each year. They employ dozens, and sometimes more than a hundred, of Ph.D. students.
In traditional academic cultures, these professors are well liked by the university president, because they bring in a lot of money.
Their Ph.D. students add them as co-authors to research papers, typically in the last position. This indicates that they are the head of the household i.e. the professorship, whether they actually contributed to the paper or not.
Maybe the complaining professor was such a heavy hitter, complaining that he could hardly be expected to check every paper his name is put on. He probably barely knows what gets published under his name and probably wouldn’t know about the content of an article either.
This is a problem, morally and formally. It is a moral problem, because scientific consensus is clearly that you should only be a co-author if you actually contributed to the paper in a non-trivial way. Providing for a conducive work environment or similar does not count.
It is a formal problem, at least in Germany, because every university president has agreed to the German National Science (DFG) foundation’s codex of good scientific practice. And this codex explicitly forbids co-authorship without non-trivial scientific contribution.
Not complying with the DFG codex bears the risk of being excluded from DFG funding. These colleagues are a risk to anyone at the same institution should the DFG eventually decide to step in and sanction non-complying universities.
Leave a Reply