A cautionary tale about the impact of AI on human design teams. (January 2021)

• Record Type: Journal Article
• Authors: Zhang, Guanglu; Raina, Ayush; Cagan, Jonathan; McComb, Christopher
• Is Part Of: Design studies. Volume 72(2021)
• Publication Date: 2021-01
• Subjects: artificial intelligence -- collaborative design -- engineering design -- human-computer interaction -- problem solving; Design -- Periodicals; 745.405
• DOI: 10.1016/j.destud.2021.100990
• Languages: English
• ISSNs: 0142-694X

Abstract

Abstract : Recent advances in artificial intelligence (AI) offer opportunities for integrating AI into human design teams. Although various AIs have been developed to aid engineering design, the impact of AI usage on human design teams has received scant research attention. This research assesses the impact of a deep learning AI on distributed human design teams through a human subject study that includes an abrupt problem change. The results demonstrate that, for this study, the AI boosts the initial performance of low-performing teams before the problem change but always hurts the performance of high-performing teams. The potential reasons behind these results are discussed and several suggestions and warnings for utilizing AI in engineering design are provided. Highlights: The AI hurts the performance of high-performing human design teams. The AI boosts the initial performance of low-performing human design teams. The AI reduces the mental demand of human designers in high-performing teams. Human designers in high-performing teams aided by AI have an illusion of success. The context and interaction of AI is critical for effectiveness.