Designerly Understanding: Information Needs for Model Transparency to Support Design Ideation for AI-Powered User Experience

UX designers lack adequate tools to understand AI as a design material: The study found that 23 UX practitioners frequently work with pre-trained models like language models but lack proper support for design ideation, creating a significant gap between AI capabilities and responsible design implementation.

Transparency frameworks need to be redesigned for designers, not just technologists: Current model transparency reporting is insufficient for UX practitioners' needs. The research identifies specific information gaps that designers require to make informed decisions about whether, where, and how to integrate AI into user experiences.

Designer understanding of AI limitations is critical for responsible AI: The study emphasizes that UX designers play a "pivotal role in Responsible AI" by serving as gatekeepers who can prevent harmful AI implementations if given proper transparency information about model limitations and behaviors.

Designers need practical, interrogation-focused information about models: Rather than technical specifications, practitioners require transparency information that helps them understand what questions to ask about AI systems and how to probe model behavior during the design process—supporting active interrogation rather than passive documentation review.

Bridging the AI-design gap requires cross-disciplinary collaboration: The research calls for transparency frameworks that speak to designers' language and concerns, suggesting that responsible AI requires not just better AI systems, but better communication between AI developers and the UX professionals who shape how people interact with these systems.