The domain-specific bias in creativity evaluation: human evaluators favor human-created art, but prefer AI outputs for divergent thinking

Abstract

How do humans perceive and evaluate creativity when it is attributed to artificial intelligence (AI)? This research examined how labeling a source as “AI” versus “human” systematically biases creativity evaluations across different domains. In experiment 1 (painting and poetry), outputs labeled as human-created were rated as more creative than identical AI-labeled outputs in artistic creation. Experiment 2, which manipulated the attribution of co-created works, revealed that the “human ideation with AI assistance” label enhanced perceived originality in poetry, whereas those attributed primarily to humans remained dominant in evaluations of painting. Experiment 3 utilized a divergent thinking task (Alternative Uses Test), found a reversal: the same output received higher creativity ratings when bearing an AI label. The findings reveal a domain-specific pattern in attitudes toward AI creativity: viewing collaboration positively yet maintaining a persistent bias in valuing AI’s creative contributions.