A measure to quantify predatory publishing is urgently needed.

Abstract

The issue of predatory publishing is of increasing concern to the academic community. In this letter, we express more concern than hope about a recently launched online machine-learning tool that identifies suspected predatory journals based on existing black/white lists and textual information from journal websites. First, the tool relies on outdated and criticized blacklists, cannot capture cloned or hijacked journals, and may misclassify legitimate journals as "suspected predatory". Second, a gray zone in predatory publishing exists where some unscholarly characteristics might exist, although the journal overall might not be considered "predatory". We tested this tool and found that it classified three well-established journals in the field of academic publishing as "suspected predatory". This may lead to undeserving negative publicity without concrete evidence of "predatory" behavior or characteristics. We argue that this tool is very premature and may lead to unfair journal classification. Considerable accountability is needed to fortify its development. We advocate for an inclusive system that involves international stakeholders, and that benefits the academic community as a "warning" system.