AI evidence extraction

Generative Artificial Intelligence in Scientific Publishing: Ethical Governance, Challenges, and Responsibilities

Dawid Gruszczyński, Kacper Nijakowski, Jowita Halupczok-Żyła, Marek Ruchała, Jarosław Walkowiak, Nadia Sawicka-Gitaj · Journal of Medical Science · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

9/10

Relevance

1/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.20883/medical.e1571

Methodology & findings

Study design

Narrative review synthesizing current evidence on AI application in scientific publishing.

Main result

The study found that "AI tools demonstrably enhance linguistic quality, streamline editorial processes, and promote global inclusivity by reducing language barriers." However, "recurrent risks were identified, including hallucinated content and references, algorithmic bias, lack of transparency ('black box' decision-making), confidentiality breaches, and misuse in peer review." Additionally, "Current policies consistently prohibit AI authorship and emphasise mandatory disclosure, yet substantial heterogeneity persists in permitted uses, reporting standards, and enforcement mechanisms across journals and publishers."

Reports effect sizes.

Research paradigm

Interpretive/critical analysis of policy and practice in scholarly publishing

Author conclusions

The authors conclude that "AI is becoming an integral component of scholarly publishing, offering meaningful benefits alongside significant risks. Responsible integration requires harmonised guidelines, transparent disclosure, rigorous human verification, and sustained editorial oversight. Preserving human judgment and accountability is crucial to ensuring research integrity and maintaining trust in the scientific record as AI technologies continue to evolve."

Risk of bias

Selection bias (narrative review without systematic search protocol); Potential publication bias (reliance on published policy documents and position statements); Lack of explicit quality assessment criteria for included studies

Open questions raised

The authors identify the need for harmonised guidelines across journals and publishers, improved enforcement mechanisms, and ongoing evolution in response to developing AI technologies. They note 'substantial heterogeneity persists in permitted uses, reporting standards, and enforcement mechanisms across journals and publishers.'
The authors identify the need for: (1) harmonised guidelines across journals and publishers regarding AI use in publishing; (2) development of transparent disclosure standards; (3) rigorous human verification protocols; (4) sustained editorial oversight mechanisms; and (5) clarification of permitted uses and reporting standards to address current heterogeneity in policies.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 78%

Explore related topics

Responsible AI Use Peer Review Literature Review