Concrete Problems for Developing Human-Centered Agents

Presented by Gagan

Gagan is part of the AI Frontiers group and co-leads research on AutoGen, a framework for building multi-agent AI systems. His work lies at the intersection of Artificial Intelligence and Human-Computer Interaction, with a focus on making AI systems more capable, interactive, and useful to people. Before joining Microsoft Research in 2022, Gagan completed his Ph.D. in Computer Science at the University of Washington, advised by Dan Weld. At UW, he was part of the Lab for Human-AI Interaction, where he studied how AI systems can complement human decision-making.

Abstract

Reflecting on his experience developing AutoGen—an open-source framework for building agents and AutoGen-based applications—this talk outlines three concrete challenges in creating human-centered agents: (1) reliably completing complex, multi-step tasks; (2) maintaining common ground between people and agents; and (3) auditing complex agent behaviors. Bansal will share demos and evaluations highlighting his progress on these challenges, as well as open opportunities for the HCI community.