Generative Artificial Intelligence in Scientific Research and Scholarly Communication: A PRISMA-Compliant Systematic Review

Muhammad Pasha

Background and Objective: Generative Artificial Intelligence (GenAI), particularly Large Language Models (LLMs), has rapidly entered scientific research and scholarly publishing workflows. While these systems offer efficiency gains in writing, peer review, and knowledge synthesis, concerns persist regarding hallucination, bias, authorship integrity, and governance. This study systematically synthesizes peer-reviewed evidence on applications, risks, ethical implications, and governance frameworks related to GenAI use in scientific research and scholarly communication. Materials and Methods: A PRISMA 2020-compliant systematic review was conducted across Scopus, Web of Science, IEEE Xplore, ACM Digital Library, and PubMed for studies published between January 2018 and March 2025. Eligible studies included empirical research, systematic reviews, experimental evaluations, and policy analyses addressing GenAI in academic research workflows. Data were extracted and thematically synthesized. Results: The 112 peer-reviewed articles met the inclusion criteria. GenAI applications were identified in manuscript drafting, literature synthesis, peer review assistance, hypothesis generation, coding, and data interpretation. Efficiency gains were consistently reported; however, risks included hallucinated citations, epistemic bias, reproducibility challenges, data privacy concerns, authorship ambiguity, and academic integrity violations. Governance responses varied across publishers and institutions, with emerging emphasis on disclosure policies, human oversight, AI literacy training, and responsible-use frameworks. Conclusion: The GenAI can enhance scholarly productivity when deployed under transparent, human-centered governance structures. Coordinated editorial, institutional, and regulatory safeguards remain essential to preserve scientific integrity and trust.

View Fulltext Back