Psychological Mechanisms Behind GenAI-Assisted Creativity Gains

Identify and quantify the psychological mechanisms that mediate gains in creative performance when humans are supported by generative AI systems such as large language models, specifically determining the relative contributions of broader ideation, deeper engagement, and heightened motivation.

Background

Most existing studies emphasize outcome comparisons (e.g., A/B experiments) rather than the processes underlying performance changes in GenAI-assisted creativity. While the meta-analysis documents modest but robust creative performance gains with GenAI support, the drivers of these gains are not established.

The paper explicitly raises uncertainty about whether improvements stem from expanded ideation breadth, increased depth of engagement due to reduced cognitive load, or greater motivation from interacting with GenAI, calling for mechanism-focused designs to disentangle these pathways.