Can generative AI improve academic journal selection decisions? Assessing traditional and algorithmic approaches in medical research.

INTRODUCTION: Journal selection is a critical step in the scientific publishing process, influencing the visibility, impact, and credibility of the published work. This task has become increasingly complex due to the proliferation of journals, predatory practices, and the diversity of editorial criteria. This narrative review presented an overview of classical tools, artificial intelligence (AI)-driven platforms, and generative models (ChatGPT, Grok) used to recommend suitable journals for an unpublished manuscript.

METHODS: Six tools were tested (Springer Journal Finder, Jane, Manuscript Matcher, Trinka Journal Finder, ChatGPT, and Grok) using either the abstract or full text of a clinical article on nonspecific low back pain. The results were compared based on thematic relevance, availability of bibliometric indicators, and transparency of the recommendations.

RESULTS: Classical tools are limited by their narrow editorial scope and the absence of key indicators. AI platforms offer broader coverage but sometimes lack precision for targeted topics. Generative tools stand out for their ability to structure recommendations, although the data provided (impact factor, fees, timelines) are often inaccurate or unverifiable. Several technological biases and algorithmic limitations impact the overall reliability of these systems.

CONCLUSION: While AI tools expedite initial journal identification, they frequently suggest journals outside the manuscript's scope and provide incorrect journal metrics. These systems function best as exploratory instruments rather than authoritative advisors. The most successful approach positions the researcher as the primary decision-maker who employs computational assistance to survey options while exercising scholarly judgment for final determinations.