Investigating Creativity in Humans and Generative AI Through Circles Exercises

TL;DR

This study investigates the prevalence of narrow creativity in both humans and Generative AI using the Circles Exercise as a design-space probe. It compares human outputs with GenAI outputs under naive and Chain-of-Thought prompting, revealing similarly constrained exploration and a heavy reliance on frequent categories and material-use patterns. While advanced prompting partially mitigates narrow creativity, it does not substantially broaden creative boundaries, indicating intrinsic limits to autonomous creativity in both humans and AI. The findings highlight opportunities to design creativity-support tools that combine human intuition with AI-generated references, using structured interaction mechanisms to promote broader design-space exploration.

Abstract

Generative AI (GenAI) is transforming the creativity process. However, as presented in this paper, GenAI encounters "narrow creativity" barriers. We observe that both humans and GenAI focus on limited subsets of the design space. We investigate this phenomenon using the "Circles Exercise," a creativity test widely used to examine the creativity of humans. Quantitative analysis reveals that humans tend to generate familiar, high-frequency ideas, while GenAI produces a larger volume of incremental innovations at a low cost. However, similar to humans, it struggles to significantly expand creative boundaries. Moreover, advanced prompting strategies, such as Chain-of-Thought (CoT) prompting, mitigate narrow creativity issues but still fall short of substantially broadening the creative scope of humans and GenAI. These findings underscore both the challenges and opportunities for advancing GenAI-powered human creativity support tools.

Figures (5)

• Figure 1: The frequency distribution of categories of objects drawn by humans and GenAI, with a more even distribution across different categories indicating better performance.
• Figure 2: Example of Drawn Object Categories: A) Human Sketched; B) GenAI-Generated, categorized into: 1) Animals, 2) Sport Equipment, 3) Foods, 4) Icons, 5) Daily Objects, and 6) Natural Elements.
• Figure 3: Example of Approaches to Material Utilization: A) Human Sketched; B) GenAI Generated, categorized into: 1) Complex Compositions 2) Circle-based Abstraction 3) Use as Background.
• Figure 4: The frequency distribution of approaches to material utilization is analyzed, with a more even distribution across different approaches indicating better performance.
• Figure 5: Example of Artistic Expression: A) Human Sketched; B) GenAI-Generated, categorized into: 1) Simple Sketches 2) Detailed Illustrations 3) Use of Color 4) Annotations and Labels.