Experimental study on improving the quality of interdisciplinary design: Collaborate with AI psychology experts for design develop process

In interdisciplinary collaboration design, arguments stemming from a lack of relevant knowledge can lead to wasted time and negative emotions. It is difficult for designers to maintain close collaboration with other experts and master diverse knowledge sufficiently within short timeframes. To address this problem, a within-subjects study was conducted based on the Double Diamond Model development process. This study developed an artificial intelligence (AI) agent based on behavioral activation clinical guidelines, self-help manuals, and interviews with 20 psychological experts. Thirty-one designers completed both the traditional design process and the AI-assisted design process as required. Sixty-two design outcomes were evaluated by design and psychology experts. The research results were analyzed based on expert ratings, AI conversation records, and open-ended questionnaires. The findings indicate that in interdisciplinary design, AI expert-assisted approaches yield design outcomes with greater effectiveness, completeness, and applicability. Analysis of conversation types and usage experiences indicates that guided interactions lead to higher-quality design outcomes, whereas command-based interactions have a negative impact. In AI-assisted interdisciplinary design, AI assists designers primarily by compensating for interdisciplinary knowledge, integrating information, optimizing details, and providing constructive improvement directions. To improve the AI-assisted design experience, AI needs to provide answers in multiple formats, connect with other design tools, and avoid excessive use of professional terms. Based on experimental and interview data, this study proposes the AI-assisted Interdisciplinary Design (AID) model, which aims to clarify the processes, roles, and capabilities of AI in interdisciplinary design.