AI evidence extraction

Best Practices for the Use of Generative Artificial Intelligence for Authors, Peer Reviewers, and Editors

K. R. Sethuraman · International Journal of Advanced Medical and Health Research · 2023

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

9/10

Relevance

0/4

Quality (LMQS)

Evidence

Citations

0.04

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.4103/ijamr.ijamr_306_23

Methodology & findings

Study design

Narrative review and synthesis of guidelines from professional organizations (COPE and World Association of Medical Editors) combined with descriptive cataloging of AI tools and their applications in research workflows..

Main result

The paper identifies that "generative AI could produce a completely fabricated scientific paper that appears authentic" and notes that "a major problem with a large language GenAI model is hallucination; it can generate fictitious information, which is then presented as an accurate fact." The authors also present best practices across multiple research domains, noting that "AI-driven search engines like Powerdrill or Litmaps can analyze relevant literature to help identify any lacuna in the current knowledge and quickly comprehend variations among published papers."

Research paradigm

Pragmatist/Applied ethics

Author conclusions

The authors conclude: "researchers and academicians can adopt the current best practices to responsibly use various generative AI tools and be more efficient and effective in conducting research and scholarly publication. Meanwhile, scientific publishers must issue guidelines to curb the misuse of GenAI tools and adopt various strategies to ensure that the authenticity of published works is maintained in the era of generative AI."

Risk of bias

Selective tool coverage - authors acknowledge listing only a subset of available AI tools; No systematic methodology for identifying or evaluating tools discussed; Potential publication bias in favor of tools that are more widely promoted; Limited evidence base - single reference cited (n=1) for this being a non-empirical review; Selection bias in choice of AI applications discussed (non-exhaustive); Potential conflict of interest: advocacy for AI tool adoption by researchers and publishers; Unverified claims about AI tool capabilities without empirical evidence

Open questions raised

The paper identifies the need for: (1) publisher guidelines to regulate GenAI tool use; (2) strategies to detect AI-generated or altered content; (3) validation and verification methods for AI-generated outputs; and (4) frameworks for responsible disclosure of AI use in scholarly work.
The paper identifies the need for scientific publishers to develop guidelines to curb misuse of GenAI tools and adopt strategies to maintain the authenticity of published works in the era of generative AI.
The paper does not explicitly identify future research directions but implies the need for: (1) development of tools to detect AI-generated or altered content by editors, (2) further guidelines by scientific publishers to curb misuse of GenAI tools, and (3) strategies to maintain authenticity of published works.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 64%

Explore related topics

AI Manuscript Writing Academic Writing Position Paper