An empirical investigation of the impact of ChatGPT on creativity
Abstract
This paper investigates the potential of ChatGPT for helping humans tackle problems that require creativity. Across five experiments, we asked participants to use ChatGPT (GPT-3.5) to generate creative ideas for various everyday and innovation-related problems, including choosing a creative gift for a teenager, making a toy, repurposing unused items and designing an innovative dining table. We found that using ChatGPT increased the creativity of the generated ideas compared with not using any technology or using a conventional Web search (Google). This effect remained robust regardless of whether the problem required consideration of many (versus few) constraints and whether it was viewed as requiring empathetic concern. Furthermore, ChatGPT was most effective at generating incrementally (versus radically) new ideas. Process evidence suggests that the positive influence of ChatGPT can be attributed to its capability to combine remotely related concepts into a cohesive form, leading to a more articulate presentation of ideas.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
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?
Several studies find benefits from GenAI in creative tasks, but these findings are typically limited to specific creative tasks.
“For example, several studies find benefits from GenAI in creative tasks, but these findings are typically limited to specific creative tasks (e.g., story writing, ideating business models) [16, 39, 64, 72].”
From Generative AI and Creativity: A Systematic Literature Review and Meta-Analysis (2025) · cites this paper as [39]
Recent studies examining human-GenAI teams versus GenAI-only outputs report no significant performance gains from adding humans to GenAI-generated content.
“recent studies examining the latter report no significant performance gains from adding humans to GenAI-generated content [39].”
From Generative AI and Creativity: A Systematic Literature Review and Meta-Analysis (2025) · cites this paper as [39]
The majority of studies identified in our literature review rely on relatively naive GenAI setups-often based on simple one-shot prompts or by providing participants with unmodified, off-the-shelf LLMs.
“The majority of studies identified in our literature review rely on relatively naive GenAI setups-often based on simple one-shot prompts (e.g., as in [37]) or by providing participants with unmodified, offthe-shelf LLMs (e.g., as in [39]).”
From Generative AI and Creativity: A Systematic Literature Review and Meta-Analysis (2025) · cites this paper as [39]
Current GenAI models often tend to recombine familiar elements across contexts rather than generating radically novel ideas, which may lead to more incremental innovation.
“Further, current GenAI models often tend to recombine familiar elements [39] across contexts rather than generating radically novel ideas, which may lead to more incremental innovation.”
From Generative AI and Creativity: A Systematic Literature Review and Meta-Analysis (2025) · cites this paper as [39]
Previous qualitative research highlighted students' fears of unfair judgement for declaring AI use.
“Previous qualitative research 19,32 highlighted students' fears of unfair judgement for declaring AI use.”
From Style, sentiment, and quality of undergraduate writing in the AI era: A cross-sectional and longitudinal analysis of 4,820 authentic empirical reports (2025) · cites this paper as [32]
The citation lists haven’t been fetched yet
References 27 · Cited by 179
- Authors
- Lee, B. C., & Chung, J.
- Affiliations
- University of Houston (US)
- Rice University (US)
- Corresponding author
- Lee, Byung Cheol
- Venue
- Nature Human Behaviour
- Published
- 2024-08-12
- Volume
- 8
- Issue
- 10
- Pages
- 1906-1914
- ISSN
- 2397-3374 (ISSN-L)
- OpenAlex
- W4401518309
- Cited by
- 179
- References
- 27
- Citations by year
- 2026: 102 · 2025: 75 · 2024: 2
- FWCI
- 73.23
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- Nature Human Behaviour (publishedVersion)
- PubMed (publishedVersion)
- RePEc: Research Papers in Economics (submittedVersion)
- OSF Preprints (OSF Preprints) (submittedVersion)
- Topics
- Social Sciences › Psychology › Experimental and Cognitive Psychology › Creativity in Education and Neuroscience
- Physical Sciences › Engineering › Mechanical Engineering › Design Education and Practice
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
- Keywords
- ChatGPT, creativity, generative AI, idea generation, problem solving, innovation, everyday problem solving, empathic concern, constraints, conceptual combination
- MeSH terms
- Adolescent, Adult, Creativity, Female, Humans, Male, Problem Solving, Internet, Young Adult
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
