Microsoft Research explores how AI failures teach valuable lessons for real-world applications

Jennifer Neville, a partner research manager at Microsoft Research, discusses how evaluating AI systems beyond traditional benchmarks reveals surprising failures that help meet user needs. Her research examines machine learning and AI for interactive domains and structured data, focusing on how training data affects AI behavior and user alignment. Neville has published over 130 papers with more than 10,000 citations and received honors including a National Science Foundation CAREER Award.