Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

How AI Ideas Affect the Creativity, Diversity, and Evolution of Human Ideas: Evidence From a Large, Dynamic Experiment

Joshua Ashkinaze, Julia Mendelsohn, Li Qiwei, Ceren Budak, Eric Gilbert

TL;DR

The study tackles how passive exposure to AI-generated ideas reshapes human creativity, diversity, and cultural evolution of ideas. It employs a large-scale, dynamic multiple-world design around the Alternate Uses Task, manipulating AI exposure and disclosure while seeding ideas from prior participants. Findings show high AI exposure increases collective diversity and its rate of change but does not enhance individual creativity; disclosure has limited main effects with heterogeneous adoption tied to self-perceived creativity and task difficulty. These results imply AI ideas can augment collective variety without boosting individual creativity, highlighting nuanced implications for AI integration in cultural and collaborative settings.

Abstract

Exposure to large language model output is rapidly increasing. How will seeing AI-generated ideas affect human ideas? We conducted an experiment (800+ participants, 40+ countries) where participants viewed creative ideas that were from ChatGPT or prior experimental participants and then brainstormed their own idea. We varied the number of AI-generated examples (none, low, or high exposure) and if the examples were labeled as 'AI' (disclosure). Our dynamic experiment design -- ideas from prior participants in an experimental condition are used as stimuli for future participants in the same experimental condition -- speaks to the interdependent process of cultural creation: creative ideas are built upon prior ideas. Hence, we capture the compounding effects of having LLMs 'in the culture loop'. We find that high AI exposure (but not low AI exposure) did not affect the creativity of individual ideas but did increase the average amount and rate of change of collective idea diversity. AI made ideas different, not better. There were no main effects of disclosure. We also found that self-reported creative people were less influenced by knowing an idea was from AI and that participants may knowingly adopt AI ideas when the task is difficult. Our findings suggest that introducing AI ideas may increase collective diversity but not individual creativity.

How AI Ideas Affect the Creativity, Diversity, and Evolution of Human Ideas: Evidence From a Large, Dynamic Experiment

TL;DR

The study tackles how passive exposure to AI-generated ideas reshapes human creativity, diversity, and cultural evolution of ideas. It employs a large-scale, dynamic multiple-world design around the Alternate Uses Task, manipulating AI exposure and disclosure while seeding ideas from prior participants. Findings show high AI exposure increases collective diversity and its rate of change but does not enhance individual creativity; disclosure has limited main effects with heterogeneous adoption tied to self-perceived creativity and task difficulty. These results imply AI ideas can augment collective variety without boosting individual creativity, highlighting nuanced implications for AI integration in cultural and collaborative settings.

Abstract

Exposure to large language model output is rapidly increasing. How will seeing AI-generated ideas affect human ideas? We conducted an experiment (800+ participants, 40+ countries) where participants viewed creative ideas that were from ChatGPT or prior experimental participants and then brainstormed their own idea. We varied the number of AI-generated examples (none, low, or high exposure) and if the examples were labeled as 'AI' (disclosure). Our dynamic experiment design -- ideas from prior participants in an experimental condition are used as stimuli for future participants in the same experimental condition -- speaks to the interdependent process of cultural creation: creative ideas are built upon prior ideas. Hence, we capture the compounding effects of having LLMs 'in the culture loop'. We find that high AI exposure (but not low AI exposure) did not affect the creativity of individual ideas but did increase the average amount and rate of change of collective idea diversity. AI made ideas different, not better. There were no main effects of disclosure. We also found that self-reported creative people were less influenced by knowing an idea was from AI and that participants may knowingly adopt AI ideas when the task is difficult. Our findings suggest that introducing AI ideas may increase collective diversity but not individual creativity.
Paper Structure (75 sections, 6 equations, 14 figures, 21 tables)

This paper contains 75 sections, 6 equations, 14 figures, 21 tables.

Table of Contents

  1. Introduction
  2. Findings
  3. Defining Concepts and Variables
  4. Creativity
  5. Idea Diversity & AI Adoption
  6. Related Work
  7. Human-AI Co-Creation
  8. Crowdsourced Creativity
  9. Collective Dynamics & Multiple-Worlds Experiments
  10. Our Contributions
  11. Experiment
  12. Summary
  13. Stimuli Creation
  14. Picking AUT Items.
  15. Model & Prompting Strategy.
  16. ...and 60 more sections

Figures (14)

  • Figure 1: Graphical depiction of experiment. The task (Panel 1) is to submit a creative idea after seeing examples, where examples are from humans or AI. We vary (Panel 2) the amount of AI ideas in the example set (exposure) and if AI ideas are labeled as such (disclosure). The experiment is dynamic (Panel 3). Responses from prior participants serve as examples for future participants.
  • Figure 2: Participants are randomized to a sequence of 5 trials. In each trial, participants generate a creative use for an item under a specific experimental condition. Neither items nor conditions repeat in a 5-trial sequence.
  • Figure 3: Study description provided to participants.
  • Figure 4: Pre-Treatment Questions
  • Figure 5: Task Instructions
  • ...and 9 more figures