Generative AI Flooding Academic Journals with Spam

Hi everyone, welcome to TechCrunch’s regular AI newsletter. This week, we’re diving into a troubling development: generative AI is now spamming academic journals.

In a recent post on Retraction Watch, assistant professors of philosophy Tomasz Żuradzk and Leszek Wroński highlighted that three journals published by Addleton Academic Publishers are composed almost entirely of AI-generated articles. These journals follow a repetitive template filled with buzzwords like “blockchain,” “metaverse,” “internet of things,” and “deep learning.” Moreover, they share the same editorial board, which includes ten deceased members, and are registered at a nondescript address in Queens, New York, that turns out to be a house.

You might wonder why this is a significant issue. Isn’t dealing with AI-generated spam a regular part of internet life? While that might be true, these fake journals highlight how easily the systems used to evaluate researchers for promotions and hiring can be manipulated. This is a worrying signal for knowledge workers in other fields as well.

For instance, on CiteScore, a widely used evaluation system, these journals rank in the top 10 for philosophy research. This is achieved through extensive cross-citation among the fake journals. Żuradzk and Wroński discovered that out of 541 citations in one Addleton journal, 208 were from other fake publications by the same publisher.

These rankings significantly impact universities and funding bodies, influencing decisions on academic awards, hiring, and promotions. As Żuradzk and Wroński point out, this manipulation could alter researchers' publication strategies, affecting the academic landscape.

One could argue that the flaw lies with CiteScore as a metric. Indeed, its vulnerability to manipulation is apparent. However, it’s also clear that the misuse of generative AI is disrupting critical systems that people's careers depend on in unexpected and harmful ways.

There’s a potential future where we rethink and redesign systems like CiteScore to be more fair, comprehensive, and inclusive. Unfortunately, the current trajectory suggests a future where generative AI continues to cause chaos, endangering professional lives and the integrity of academic research.

We urgently need to address this issue before it causes more damage. Let’s hope for a swift course correction.