AI evidence extraction

Citations to ChatGPT: A Cited Reference Analysis Across Disciplines

Robert Tomaszewski · Science & Technology Libraries · 2026

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

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.1080/0194262x.2026.2655254

Methodology & findings

Study design

Bibliometric citation analysis using Scopus database search.

Sample

N = 2963, 3 groups

Primary method

Bibliometric analysis methods including keyword co-occurrence network analysis using VOSviewer software. Comparative analysis of citation visibility across different LLMs (Claude, Ernie Bot, Gemini, and LLaMA). Analysis of geographic trends, institutional affiliations, source diversity, disciplinary distributions, and thematic structures.

Main result

The study found that "Results show rapid growth in ChatGPT citations since late 2022, led by computer science, the social sciences, and healthcare. The United States and China account for the largest share of publications. ChatGPT demonstrates substantially greater citation visibility than other LLMs, reflecting early adoption, accessibility, and interdisciplinary integration."

Reports effect sizes.

Research paradigm

Positivist/Empiricist

Author conclusions

The authors conclude that "Persistent inconsistencies in AI attribution and disclosure practices highlight the need for clearer citation standards and research policy guidance." They emphasize that "By focusing on formal citations rather than informal mentions, this study provides a more precise account of how generative AI tools enter the scholarly record."

Risk of bias

Selection bias: Study limited to formal citations in Scopus, potentially missing informal mentions and unpublished uses; Database bias: Scopus coverage varies by discipline and may underrepresent certain fields; Temporal bias: ChatGPT's rapid growth may skew toward recent publications; Language bias: Likely predominantly English-language publications in Scopus; Publication bias: Only published works captured, not grey literature; Selection bias: Database selection (Scopus only) may exclude publications in other indexing systems; Selection bias: Restriction to explicit citations may undercount informal mentions; Attribution bias: Inconsistencies in AI attribution and disclosure practices noted in results; Database selection bias - limited to Scopus-indexed publications; Selection bias - refined to specific document types (articles, conference papers, reviews), excluding other forms of scholarship; Citation practice variation - inconsistent AI attribution and disclosure practices across disciplines may affect visibility; Temporal bias - rapid evolution of ChatGPT usage and competing LLMs between 2021-2025 may skew trends

Limitations

The study focused "on formal citations rather than informal mentions," which means it "provides a more precise account of how generative AI tools enter the scholarly record" but may not capture the full extent of ChatGPT's influence in scholarly work
The abstract notes "Persistent inconsistencies in AI attribution and disclosure practices" which suggests limitations in the citation data itself.

Open questions raised

The study identifies the need for clearer citation standards and research policy guidance for AI tools. It suggests that inconsistencies in AI attribution and disclosure practices require standardization across disciplines.
The authors identify the need for "clearer citation standards and research policy guidance" regarding AI tool attribution and disclosure in scholarly research.
The study identifies the need for clearer citation standards and research policy guidance for AI tools in scholarly research. The authors note that inconsistencies in how ChatGPT and other LLMs are cited and acknowledged require standardization.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 59%

Explore related topics

Citation Integrity Research Integrity