[1H5-OS-5b] OS-5

5:40 PM - 5:55 PM JST(8:40 AM - 8:55 AM UTC)

[1H5-OS-5b-01] Collective Creativity in Human–AI Hybrid Societies

〇Shota Shiiku^1,2,6, Raja Marjieh³, Manuel Anglada-Tort^4,6, Nori Jacoby^5,6 (1. Shizuoka University, 2. ExaWizards Inc., 3. Princeton University, 4. Goldsmiths, University of London, 5. Cornell University, 6. Max Planck Institute for Empirical Aesthetics)

Keywords:

Human-AI Interaction,Collective Intelligence,Collective Creativity,Social Networks,Large Language Models

Rapid advances in machine learning and generative AI are giving rise to hybrid societies, in which ideas, beliefs, and creative artefacts evolve through continuous interaction between humans and machines. A central open question is how the introduction of AI agents into human social systems alters trade-offs such as those between creativity and diversity, yet these influences are difficult to capture using traditional methods. This study investigates how human–AI interactions shape collective creativity using social network experiments. Participants selected, modified, and transmitted stories generated by their network neighbours, allowing us to observe how micro-level interactions scale up to collective dynamics. Hybrid networks fundamentally alter the convergence dynamics observed in purely human or purely AI networks: humans embedded in hybrid networks became substantially more creative, while AI agents maintained higher levels of collective diversity when interacting with humans. These findings provide causal empirical evidence that AI influences human creativity not only through direct dyadic interaction, but also through emergent, network-level processes.