Fake no more: The redemption of ChatGPT in literature reviews

Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F60460709%3A41110%2F25%3A102884" target="_blank" >RIV/60460709:41110/25:102884 - isvavai.cz</a>
Result on the web
<a href="https://doi.org/10.1080/08989621.2025.2465619" target="_blank" >https://doi.org/10.1080/08989621.2025.2465619</a>
DOI - Digital Object Identifier
<a href="http://dx.doi.org/10.1080/08989621.2025.2465619" target="_blank" >10.1080/08989621.2025.2465619</a>

Result language
angličtina
Original language name
Fake no more: The redemption of ChatGPT in literature reviews
Original language description
This paper revisits earlier concerns regarding the reliability of ChatGPT for conducting academic literature reviews, particularly its tendency to generate fabricated references and invalid DOIs. Building on prior critical assessments, the authors evaluate recent advances in AI-assisted research through an empirical test of OpenAI’s Deep Research tool. Using a previously published prompt, they assess the accuracy, validity, and scholarly relevance of AI-generated references in the medical field. The findings demonstrate a marked improvement in citation accuracy, DOI validity, and disciplinary relevance, with the identified articles showing high citation impact. These results suggest that recent developments significantly enhance the credibility of AI tools for literature reviews. The paper concludes that, contrary to earlier recommendations, ChatGPT equipped with Deep Research can now be responsibly integrated into academic research workflows, while also raising broader questions about the future of scholarship.
Czech name
—
Czech description
—

Type
J<sub>imp</sub> - Article in a specialist periodical, which is included in the Web of Science database
CEP classification
—
OECD FORD branch
30311 - Medical ethics

Publication year
2025
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Similar results(10)