CoChat
Sign Up
Cited by 165Creativity in Education and NeuroscienceArtificial Intelligence ApplicationsArtificial Intelligence in Games

The originality of machines: AI takes the Torrance test

Open access
Guzik, Erik E., Byrge, Christian, Gilde, Christian·2023·Journal of Creativity
DOI 10.1016/j.yjoc.2023.100065
Open this paper in CoChat — free account

Abstract

This exploratory research investigated the creative abilities of OpenAI's large language model, ChatGPT, based on the GPT-4 architecture, as assessed by the Torrance Tests of Creative Thinking. In comparison to human samples and a national percentile from Scholastic Testing Services, ChatGPT's performance was analyzed for fluency, flexibility, and originality. Results indicated that ChatGPT scored within the top 1% for originality and fluency, and showed high scores for flexibility, thus highlighting the current creative abilities of AI and the potential of AI systems to support and augment human creativity in new and meaningful ways. The study encourages additional research to further define, measure, and develop creativity in the era of advanced AI.

Integrity signals

Registry

Crossref · 10.1016/j.yjoc.2023.100065

Cites

22 of 22 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

We couldn’t retrieve this paper’s full text, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Some studies show GenAI outperforming humans on creative tasks.

    “Other studies show GenAI outperforming humans. Consultants using GenAI [30] produced more than 40% higher quality solutions in a controlled study, where quality was an aggregate score comprising creativity, analytical thinking, writing proficiency, and persuasiveness.”

    From The Impact of Generative AI on Creativity in Software Development: A Research Agenda (2025) · cites this paper as [52]

  • LLM systems perform as well as or even better than humans on commonly accepted creative ideation benchmarks and on benchmarks for ideation in specific contexts.

    “The first is that LLM systems perform as well as or even better than humans on commonly accepted creative ideation benchmarks [8] and benchmarks for ideation in specific contexts [7].”

    From The Role of Human Agency in Human-AI Co-Creativity (2026) · cites this paper as [8]

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Guzik, E. E., Byrge, C., & Gilde, C.
Affiliations
  • University of Montana (US)
  • Vilnius University (LT)
  • University of Montana Western (US)
Corresponding author
Guzik, Erik E.
ORCID
Venue
Journal of Creativity
Publisher
Elsevier BV
Published
2023-12
Volume
33
Issue
3
Pages
100065
ISSN
2713-3745 (ISSN-L)
DOI
10.1016/j.yjoc.2023.100065
OpenAlex
W4386052615
Cited by
165
References
22
Citations by year
2026: 31 · 2025: 71 · 2024: 56 · 2023: 6
FWCI
39.62
Type
journal article
Language
en
Open access
Licence
Locations
  • Journal of Creativity (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Creativity in Education and Neuroscience
  • Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
  • Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence in Games
Keywords
ChatGPT, large language models, Torrance Tests of Creative Thinking, computational creativity, originality, fluency, flexibility, human-AI comparison

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review