Researchers from the University of Colorado Boulder have developed an artificial intelligence tool to identify scientific journals potentially operating under questionable publishing standards. Published in Science Advances, the study addresses the proliferation of so-called 'predatory' journals—publications that charge authors fees without providing standard peer review or editorial services.
Led by Daniel Acuña, the project leverages data from the Directory of Open Access Journals (DOAJ) to train its model in detecting red flags on journal websites, such as grammatical errors, excessive self-citations, unusual publication volumes, and lack of peer review policy disclosures.
The AI screened nearly 15,200 open-access journals and flagged over 1,400 as potentially problematic. After expert review, around 1,000 were deemed to raise credible concerns. The system, designed to be interpretable, is not intended to replace human judgment but to assist institutions and publishers in early identification.
The researchers emphasize that while flawed science can compromise knowledge foundations, tools like this one could serve as a protective mechanism or 'firewall for science.' Although the tool is not publicly available yet, the team plans to make it accessible to academic and publishing stakeholders.
Click here to read the original press release.
More News in this Theme
