Human diversity fuels collective creativity that large language models cannot simulate or sustain

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Mengchen Dong, Hiromu Yakura
• Venue: arXiv: cs.HC
• Published: 2026-07-29
• Collected: 2026-07-30
• URL: http://arxiv.org/abs/2607.26899v1
• PDF: https://arxiv.org/pdf/2607.26899v1
• DOI: UNKNOWN

Plain-Language Brief

Researchers found that diversity among human groups contributes to collective creativity, and this can't be replicated by current AI models. The study compared human creators using their native languages (L1) and non-native languages (L2) with AI-assisted creativity. L2 writers showed more diversity than L1 writers, but AI tended to reduce overall diversity. However, when individual writers used AI for ideation or refinement, it boosted their personal ratings. The study suggests that human diversity remains a valuable resource that isn't fully replicable by AI.

What the Document Says

• Diverse human groups contribute to collective creativity.
• AI challenges the engine of human creativity in two ways: it may homogenize what diverse people create, and it may replace people altogether.
• The study tested these challenges using a creative metaphor experiment with English writers (L1 and L2) without AI, with AI-generated ideas (AI ideation), or with AI refining their own ideas (AI refinement).
• L2 writers showed more collective diversity than L1 writers.
• Native-language ideation showed the most diverse pools.
• AI ideation compressed collective diversity for everyone and left the L2 advantage undetectable.
• AI refinement preserved both human diversity and individual contributions.
• The study simulated the entire writer pool using personas built from participants' real backgrounds, three model families, native-language prompting, and elevated sampling temperatures. All simulated pools fell below human pools and pushing models further induced diversity only through degenerate text.
• At the individual level, AI ideation raised writers' ratings, but this was not beneficial for collective creativity unless L2 writers used their native language.
• The study concludes that human diversity remains a valuable creative resource that current AI cannot simulate or sustain.

Key Concepts Mentioned

• human diversity
• creative metaphor experiment
• AI ideation
• AI refinement
• native-language prompting
• elevated sampling temperatures
• degenerate text

Claims Made by the Document

• Diverse human groups contribute to collective creativity.
• AI challenges the engine of human creativity.
• L2 writers show more collective diversity than L1 writers.

Evidence Used

• Diverse human groups produce diverse ideas, the raw material of innovation.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Diverse human groups produce diverse ideas, the raw material of innovation. Generative AI challenges this engine twice over: everyday AI assistance may homogenize what diverse people create, and AI-simulated diversity may replace the people altogether. We tested both challenges in a preregistered creative metaphor experiment with native (L1) and non-native (L2) English writers, who wrote without AI, with AI-generated ideas (AI ideation), or with AI refining their own ideas (AI refinement). L2 writers contributed more collective diversity than L1 writers, with native-language ideation showing the most diverse pools. AI ideation compressed collective diversity for everyone and left the L2 advantage undetectable, whereas AI refinement preserved both. We then simulated the entire writer pool using personas built from participants' real backgrounds, three model families, native-language prompting, and elevated sampling temperatures. Every simulated pool fell below every human pool, and pushing models further induced diversity only through degenerate text. However, at the individual level, AI ideation raised writers' ratings, pitting private incentives against the collective good, except when L2 writers used their native language, which benefited both. Human diversity remains a valuable creative resource that current AI cannot simulate or sustain; the design of human-AI collaborative workflows determines whether it survives.