AI evidence extraction

Transparency in research: An analysis of ChatGPT usage acknowledgment by authors across disciplines and geographies

Raghu Raman · Accountability in 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

1.17

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.1080/08989621.2023.2273377

Methodology & findings

Study design

Systematic bibliometric analysis of publications in the Dimensions database.

Sample

N = 1226, 3 groups

Primary method

Bibliometric analysis using temporal trend analysis and distribution analysis across disciplines and geographic regions. Specific statistical software and methods are not detailed in the abstract provided.

Main result

The study found that "U.S.-based authors lead in acknowledgments, with notable contributions from China and India" and that "Biomedical and Clinical Sciences, as well as Information and Computing Sciences, are engaging with these AI tools," with publications like "The Lancet Digital Health" and platforms such as "bioRxiv" being recurrent venues for such acknowledgments, indicating AI's growing impact on research dissemination.

Reports effect sizes.

Research paradigm

Positivist/empiricist

Author conclusions

"This study represents the inaugural comprehensive empirical assessment of AI acknowledgment patterns in academic contexts, addressing a previously unexplored aspect of scholarly communication," with findings being "beneficial for stakeholders, providing a basis for policy and scholarly discourse on ethical AI use in academia."

Risk of bias

Database selection bias (limited to Dimensions database only); Temporal bias (only covers November 2022 to July 2023); Publication bias (may miss unpublished or grey literature); Geographic bias (emphasis on certain regions may reflect database coverage rather than actual usage patterns); Database selection bias: reliance on single database (Dimensions) may exclude other indexing platforms; Publication venue bias: analysis limited to indexed publications, excluding grey literature; Temporal selection bias: analysis window (November 2022 - July 2023) may not capture full adoption patterns; Quality assessment not conducted: unable to evaluate acknowledgment rigor or sincerity; Database selection bias (Dimensions database only, excluding other indexing services and grey literature); Potential language bias (may exclude non-English publications); Temporal bias (limited to November 2022 to July 2023 period); Publication bias (published works only, excludes unpublished research)

Limitations

"The analysis is confined to the Dimensions database, thus potentially overlooking other sources and grey literature
Additionally, the study abstains from examining the acknowledgments' quality or ethical considerations."

Open questions raised

The study identifies the need for examination of acknowledgment quality and ethical considerations in AI usage by researchers. It also highlights the gap in understanding AI acknowledgment patterns in academic contexts prior to this investigation.
Quality and ethical considerations in AI acknowledgments; examination of grey literature and non-indexed sources; broader temporal assessment beyond the initial adoption period; detailed analysis of acknowledgment practices across specific disciplines.
The authors identify the need for future research examining the quality and ethical considerations of AI acknowledgments in scholarly work, and the potential for expanding analysis beyond the Dimensions database to include grey literature and other sources.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 61%

Explore related topics

Citation Integrity Research Integrity Content Analysis