AI is useful at the decision point—if its reasoning stays visible
A DRS2026 experiment with 40 graduate design students compared different AI support modes for selecting final concepts; AI scoring produced the strongest reported satisfaction and lowest cognitive load, while participants still wanted transparent logic.
Graduate design students compared four AI-assisted ways to select a final concept.
The Story
A Design Research Society study compared four AI support modes—textual description, pros-and-cons comparison, AI scoring, and direct recommendation—in a design selection task with 40 graduate design students. The study reported the strongest satisfaction and lowest cognitive load for AI scoring. ([dl.designresearchsociety.org](https://dl.designresearchsociety.org/drs-conference-papers/drs2026/researchpapers/220/?utm_source=openai))
The more important finding is the qualification: designers valued efficiency but still wanted to understand how the system reached its judgment. That points toward AI as a structured critique partner, not an invisible taste-maker.
A useful Studio Wensday workflow would expose the rubric. Ask the system to score concepts against explicit criteria such as distinctiveness, accessibility, brand fit, and feasibility—then keep the final decision human, contextual, and discussable.
Studio Wensday Angle
AI-assisted convergence works best when the criteria are visible and the judgment remains contestable.