Nobox

Tiffany J. Hopkins is a human-AI interaction researcher who helps organizations understand what people need before deciding what to build. Her work combines qualitative and quantitative methods to investigate how users interpret systems and how models behave in real-world settings. She translates human context into behavioral taxonomies, scenarios, rubrics, and evaluations that guide product and model decisions, including questions of helpfulness, autonomy, memory, and collaboration.

Hopkins’s approach, which she calls relational intelligence, focuses on what data means for the people represented in it. Her background spans cognitive science at UC San Diego, work at MTV, business school, and research practice at IDEO, as well as projects with organizations worldwide through Nobox. Her current research includes co-designing an evidence-grounded report-drafting system with a clinician, translating professional terminology and judgment into system rules and human-reviewed evaluations. These methods are intended to identify plausible but incorrect outputs and assess whether a system genuinely helps. Hopkins offers applied research and evaluation expertise to teams shaping human-centered products and AI systems.