Abstract
Purpose: Nowadays, many studies discuss scholarly publishing and associated challenges, but the problem of hijacked journals has been neglected. Hijacked journals are cloned websites that mimic original journals but are managed by cybercriminals. The present study uses a topic modeling approach to analyze published papers in hijacked versions of medical journals. Methods: A total of 3,384 papers were downloaded from 21 hijacked journals in the medical domain and analyzed by topic modeling algorithm. Results: Results indicate that hijacked versions of medical journals are published in most fields of the medical domain and typically respect the primary domain of the original journal. Conclusion: The academic world is faced with the third-generation of hijacked journals, and their detection may be more complex than common ones. The usage of artificial intelligence can be a powerful tool to deal with the phenomenon.