Resources to counteract AI-driven fraudulent publishing

Ascertain whether the scientific community has sufficient resources to effectively counteract the exacerbation of fraudulent scientific publishing caused by artificial intelligence and the resulting multiplication of cheating options.

Background

The paper’s closing section discusses how AI tools, while useful for copy-editing, can be exploited by paper mills and bad actors to fabricate or manipulate research content. As AI-generated texts become more convincing, traditional screening methods struggle, prompting publishers to invest in new integrity software and detection strategies.

The authors emphasize an escalating arms race between generative models and detection tools, alongside evolving guidelines from publishers and funding agencies. Against this backdrop, they explicitly state uncertainty about whether the scientific community has the necessary resources to counteract AI’s exacerbation of fraudulent publishing.