AI Interaction and Learning

The AI Interaction and Learning (AIIL) group in Microsoft Research studies how to build successful AI systems that help people accomplish more valuable work. We focus on AI systems that operate in realistic, long-horizon workflows, where users, tasks, and system behavior evolve through interaction.

A central theme of our work is closing the loop between understanding real-world AI use and building better AI systems. We study how people use AI in practice, identify limitations and opportunities, and use those insights to guide evaluation, simulation, and learning.

Our current research spans three closely connected areas:

Understanding AI Use and Evaluating Realistic Workflows: We analyze large-scale interaction and work traces to understand how people and organizations use AI, how AI is changing work practices, and where new capabilities could be most valuable. We also develop measurements and evaluations that expose how AI assistants and agents behave in realistic, evolving tasks.

User Simulation and Interactive Environments: We build models and environments that represent users, tasks, and multi-step interactions. These enable controlled study of realistic workflows and provide scalable settings for evaluating and improving AI systems.

Learning, Adaptation, and Memory: We develop methods that enable AI systems to improve through interaction and experience. Our work includes reinforcement learning and post-training, personalization and preference learning, alignment, and memory mechanisms that help agents retain and use knowledge about users, tasks, and processes over time.

The group brings together expertise in artificial intelligence, machine learning, natural language processing, human-computer interaction, recommender systems, information retrieval, and computational social science. This breadth allows us to connect behavioral understanding of real-world AI use with new environments, measurements, and learning methods.

Together with collaborators across Microsoft Research and our product partners, our work has contributed to AI systems and experiences across Microsoft, including Copilot, Office, Bing, and Xbox.

We are always interested in working with researchers who want to push the behavioral and algorithmic frontiers of AI. We value people who can think broadly about important problems, go deeply on the science, and work collaboratively across disciplines. We invite you to learn more about our research on our individual profile pages.