Leveling the Playing Field: AI-Augmented Design and the Expertise Bias in Evaluations of Creative Output

• In Event: Artificial Intelligence, Algorithms, and Creativity

Abstract

Prior research demonstrates that the expertise of a designer can bias evaluators’ assessments of the designer’s creative output. Evaluators tend to more favorably evaluate output produced by designers with perceived expertise, irrespective of the actual creativity of their output (hereafter, the expertise bias). Reporting the results of multiple experiments, we find that the expertise bias is mitigated when Artificial Intelligence (AI)’s capability to augment the creative design process is salient to evaluators, consistent with theory that AI reduces the perceived exclusivity of a designer’s knowledge which mitigates the expertise bias. However, we also demonstrate that the expertise bias is exacerbated when designers refuse to incorporate input from AI systems in their designs, consistent with theory that nonconforming behavior can enhance the perceived exclusivity of a designer’s knowledge. Collectively, our results complement prior research that explores the impact of AI in the design process by highlighting its role in mitigating biased evaluations of creative output. Moreover, our results contribute to a better understanding of practitioner surveys suggesting that some highly-experienced creative designers in practice refuse to incorporate AI systems into their creative design.

Authors

• Jordan Samet, University of Illinois Urbana-Champaign

• Michael Glenn Williamson, University of Illinois at Urbana-Champaign

• Michael Yip, University of Illinois at Urbana-Champaign