🧵 Les éditeurs de publications scientifiques s'inquiètent du rôle de l' #IA dans l'augmentation sans précédent des textes de mauvaise qualité.
Récemment, un #éditorial de la revue spécialisée Sciences mentionnait: « #AI-generated content floods #literature with poor-quality publications, casts doubt on metrics, Science and Retraction Watch investigation finds » #Science #Méthodologie #Enquête
https://www.science.org/content/article/shoddy-commentaries-quick-and-dirty-route-higher-impact-numbers-are-rise
=> More informations about this toot | More toots from fdastous@tooting.ch
🧵 Ce problème se manifeste sous la forme d'une augmentation du nombre de demandes de publications pour des textes de type : commentaire scientifique.
« Neurosurgical Review.(...) The publication had been knocked off its feet by an “unprecedented increase” in submitted commentaries that appeared to be “driven by” advances in artificial intelligence (AI) tools such as ChatGPT, he explained. »🧵
=> More informations about this toot | More toots from fdastous@tooting.ch
🧵 Ce problème est reliés à une sorte de voie de facilité.
« For authors, commentaries can be a quick and easy way to amass publications and citations. Authors “just want a PubMed-indexed article. That’s it,” says Shirish Rao, a recent medical graduate who works at a hospital in Mumbai, India. Commentaries are an ideal avenue because “you don’t really need original data,” so AI tools can generate them in almost no time »
=> More informations about this toot | More toots from fdastous@tooting.ch This content has been proxied by September (ba2dc).Proxy Information
text/gemini