An AI tool has identified over 1,000 potentially problematic open-access journals, approximately 7% of 15,000 titles screened, a scale of questionable publishing previously undetected by human watchdogs, according to Science. This revelation exposes a significant, underestimated infiltration of deceptive practices within scholarly communication.
While human-led efforts to identify problematic journals are increasing, this AI tool has uncovered a much larger, hidden network of questionable publications, some even owned by reputable publishers. The academic community now faces a reckoning with the true scale of predatory publishing, demanding a shift towards sophisticated, AI-driven verification and a re-evaluation of publishing standards.
The Anatomy of a Problematic Journal
The International Journal of Advances in Signal and Image Sciences exemplifies the tactics of problematic journals. It published 19 papers in 2024, surging to 153 in 2025, with 53% of 2025 papers featuring an author from India, according to Nature. Despite its delisting from Scopus in 2025, the journal continues to falsely advertise Scopus indexing on its website, charging up to US$1,200 per article. Delisting alone fails to halt predatory operations; these journals continue to profit from deception and a lack of enforcement.
AI-Driven Detection vs. Human Oversight
Data scientist Achal Agrawal's online tool, Journal Trends, helps researchers identify potentially problematic academic journals, according to Nature. Crucially, none of the over 1,000 journals flagged by this AI had appeared on any human-curated watchlist; some are even owned by large, reputable publishers. A critical advantage is offered by the AI-powered tool, detecting sophisticated, hidden problematic journals that traditional human-led watchlists have consistently missed, even within established publishing houses.
In contrast, human-led initiatives, such as the Directory of Open Access Journals (DOAJ), investigated 473 journals in 2024, a 40% increase from 2021, according to nature.com. While such efforts intensify, the sheer volume and hidden nature of problematic journals reveal that traditional methods struggle to keep pace. The academic community must acknowledge that its current human-led oversight mechanisms are catastrophically failing, leaving research integrity vulnerable to widespread, sophisticated exploitation.
Implications for Researchers and Publishers
The AI's findings reveal that some problematic journals are owned by large, reputable publishers, implying that even established academic institutions are either complicit in or dangerously unaware of the deep infiltration of predatory practices within their portfolios. The emergence of tools like Journal Trends will force a re-evaluation of current journal vetting and demand greater transparency from publishers. By 2026, academic institutions should integrate AI-driven verification into their publication policies to safeguard research integrity.
Without such proactive integration, the integrity of scholarly publishing will likely remain compromised by widespread, undetected predatory practices.
