CoChat
Sign Up
PreprintCited by 174Creativity in Education and NeuroscienceDesign Education and PracticeArtificial Intelligence Applications

AI and Its Impact on Creativity and Diversity: An Empirical Study of LLM-Generated Product Ideas

Open access
Terwiesch, Christian, Meincke, Lennart, Girotra, Karan, Mollick, Ethan R., Nave, Gideon, Ulrich, Karl T.·2023·SSRN Electronic Journal
DOI 10.2139/ssrn.4526071
Open this paper in CoChat — free account

Abstract

This research examines how well large language models (LLMs) generate new product ideas, focusing on new products for college students priced under $50. Through a series of studies, we identify key strengths and weaknesses of using LLMs for product innovation. Our first study shows that LLM-generated product ideas have higher average quality than human ideas (based on purchase intent) and are 7 times more likely to rank in the top 10%. Our second study demonstrates that this AI-induced creativity boost is not explained by the LLM’s more persuasive pitching skills. Our third and fourth studies identify an interesting weakness of using LLMs for brainstorming, showing that AI-generated ideas are less novel (at the idea level) and less diverse (at the set level). In our fifth study, we investigate the generalizability of these findings by analyzing prior LLM-based creativity studies and find consistently lower idea diversity across all of them. Our sixth and seventh studies investigate the efficacy of various techniques to mitigate this diversity loss. Specifically, we compare different LLMs (different vendors and different versions) and find that the more recent models are capable of generating more diverse ideas, though still falling short of achieving human-level diversity. In addition, we demonstrate the effectiveness of several techniques that can be used to increase idea diversity almost to the level of human idea generation: pooling ideas across vendors, prompt engineering (including Chain-of-Thought prompting and injecting heterogeneous personas or constraints) and creative agents that broadly explore the solution landscape with the intent of restoring diversity. Finally, in our eighth study, we show that exploiting the near-zero marginal cost of AI idea generation by scaling the number of ideas steadily improves coverage of the idea space, approaching human-level coverage. We conclude our work by presenting a set of actionable recommendations targeted to managers of innovation who want to identify better new product ideas with the help of LLMs. Previous versions of the paper were available under "Ideas are Dimes a Dozen: Large Language Models for Idea Generation in Innovation")

Integrity signals

Registry

Crossref · 10.2139/ssrn.4526071

Cites

49 of 49 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?

  • A great deal of empirical research testing the creative capabilities of LLMs has not used expert evaluators, instead using lay evaluations.

    “Specifically, a great deal of empirical research that has sought to test the creative capabilities of LLMs have not used “expert” evaluators (e.g., domain experts evaluate creative outcomes). Instead, this body of work has made use of “lay” evaluations of creativity (laypeople evaluate creative outcomes) to substantiate claims of creative generative AI7,8,15,16.”

    From Establishing the importance of co-creation and self-efficacy in creative collaboration with artificial intelligence (2024) · cites this paper as [7]

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Terwiesch, C., Meincke, L., Girotra, K., Mollick, E. R., Nave, G., & Ulrich, K. T.
Affiliations
  • Cornell University (US)
  • University of Pennsylvania (US)
Corresponding author
Ulrich, Karl T.
ORCID
Venue
SSRN Electronic Journal
Publisher
Elsevier BV
Published
2023
ISSN
1556-5068 (ISSN-L)
DOI
10.2139/ssrn.4526071
OpenAlex
W4385447813
Semantic Scholar
ed7c053b38074003bfb947c78373f69598b960b8
Cited by
174
References
49
Citations by year
2026: 41 · 2025: 70 · 2024: 51 · 2023: 12
Type
posted content
Language
en
Open access
Locations
SSRN Electronic Journal (acceptedVersion)
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
large language models, idea diversity, idea novelty, new product innovation, brainstorming, prompt engineering, chain-of-thought prompting, creative agents, idea generation, university students, low-cost products, innovation management

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