Design Theater: Benchmark Reveals Gaps in Generative UI Rationales

A new study from arXiv (2607.22928) introduces the concept of "Design Theater," where generative UI tools produce plausible design rationales that are not reflected in the actual interfaces. The researchers created a benchmark with 24 UI generation tasks covering structural, styling, and functional requirements. They evaluated 120 interfaces from five generative UI tools and found that, on average, over 25% of user-facing design rationales are not implemented. The study highlights a disconnect between stated design choices and actual implementation, raising concerns about the reliability of these tools for democratizing UI design.