AI evidence extraction

ChatGPT as an “author”: Bibliometric analysis to assess the validity of authorship

Serhii Nazarovets, Jaime A. Teixeira da Silva · Accountability in Research · 2024

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

2.98

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.1080/08989621.2024.2345713

Methodology & findings

Study design

Comprehensive literature review using bibliometric databases (Web of Science and Scopus) to identify and analyze instances where ChatGPT was credited as an author, co-author, or group author on academic papers..

Sample

N = 14, 3 groups

Primary method

Bibliometric analysis using Web of Science and Scopus databases; descriptive enumeration of cases. No inferential statistical tests mentioned in abstract.

Main result

The study identified that "Our search identified 14 papers featuring ChatGPT in such roles. In four of those papers, ChatGPT was listed as an 'author' alongside the journal's editor or editor-in-chief. Several of the ChatGPT-authored papers have accrued dozens, even hundreds of citations according to Scopus, Web of Science, and Google Scholar." This reveals widespread violation of ethical authorship guidelines despite institutional policies prohibiting such practices.

Reports effect sizes.

Research paradigm

Positivist/empiricist

Author conclusions

The authors conclude that "The findings suggest a need for corrective measures to address these discrepancies. Immediate review and amendment of the listed papers is advised, highlighting a significant oversight in the enforcement of ethical standards in academic publishing." This indicates widespread failure to enforce existing ethical guidelines regarding AI authorship.

Risk of bias

Selection bias: Search strategy limited to specific databases (Web of Science, Scopus); Database coverage bias: Papers indexed in non-English databases or preprint servers may not be captured; Citation bias: Papers with higher citation counts may be more visible and subject to detection bias; Selection bias: Database-dependent search may miss papers in non-indexed journals or preprints; Potential publication bias if only indexed papers were included; Lack of detailed inclusion/exclusion criteria explicitly stated in abstract; No information on inter-rater reliability for paper classification; Selection bias: Database coverage limitations (only Web of Science and Scopus searched); Potential publication bias: Only published papers identified; Search strategy bias: Dependent on search terms and indexing practices

Open questions raised

The authors identify the need for enforcement mechanisms and corrective measures to address the gap between ethical policy and academic practice regarding AI authorship.
The study identifies a disconnect between ethical policy and academic practice regarding AI authorship, suggesting future research into enforcement mechanisms and institutional compliance with ethical guidelines in academic publishing.
The study identifies a gap between ethical policy and academic practice regarding AI authorship, suggesting need for stronger enforcement mechanisms of existing ethical guidelines from ICMJE and COPE.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 69%

Explore related topics

Authorship & Attribution Research Integrity Literature Review