GPT-3 fabricates plausible-looking but fake references and apparently invents the research studies they cite¶
Claim · All claims
Evidence · 1 study · theoreticalr2·q1
Currency · newest study 2022 · checked against its sources 2026-10-28 (automated, 1 of 1 entry) · not reviewed by a person · overdue for a recheck since 2025: tool and model accuracy changes quickly
Subclaims¶
Reading the evidence codes
q is the design tier, i the impact, n the sample, and the kind with its r
the rigour judged within that kind: what each code means.
This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.
q1 i? The GPT-3 essay cited a Dunn & Dunn (1997) article in a journal issue that does not exist, and no study by Dunn and Dunn making the claimed finding could be found. → Mike Sharples 2022
Evidence¶
Mike Sharples 2022¶
Mike Sharples. (2022). Automated Essay Writing: An AIED Opinion. International Journal of Artificial Intelligence in Education. https://doi.org/10.1007/s40593-022-00300-7
q1 · i? · theoretical · r2
The author's inspection of the generated essay's reference list: the citation names a real journal but a non-existent issue, and the author reports being unable to find any study by Dunn and Dunn claiming learning styles are flexible.
"There is a journal named Research in Education, but no issue 7(2) in 1997. Dunn & Dunn did publish research on learning styles, but not in that journal. GPT-3 has fashioned a plausible-looking but fake reference."
Discussion¶
Related Claims¶
- AI-generated ethics exercise materials carried external citations that did not survive checking, ranging from misattributed to apparently fabricated — related
- Most of the article's body text was generated by ChatGPT from author prompts, and all references ChatGPT inserted were fictional — possibly the same claim (merge candidate)
- ChatGPT fabricates references: the authors caught it inventing a non-existent 'Ribeiro and Vala, 2020' article with full bibliographic details — possibly the same claim (merge candidate)
- Review-attributed evidence: fabricated model-generated references clustered in disciplinary journals in a student verification exercise — related
- Of 30 verified hallucinated references, 13 appeared entirely fabricated and 17 were hybrid references combining a real title with fabricated authorship or metadata — related
- AI-assisted scientific writing assignments risk fabricated references and altered chemical meaning — related
- The main concern in all critical evaluations was the veracity of ChatGPT-provided references, which often do not exist — related
- The most frequently cited ethical concerns about GenAI in research were hallucinated facts, fabricated references, and misleading content — related
- GPT-3 can generate a plausible student essay from a short prompt, including a heading and citation, on its first attempt — related
Page metadata
| Field | Value |
|---|---|
status |
review |
generated |
by process:wiki-ingest on 2026-10-28 |
trust |
unverified — no one has confirmed this page's content |
Frontmatter is the page's machine-readable record — see the Schema & Guide.