Short abstract

I'll discuss three projects related to understanding how people and AI-infused systems can and should interact. In the first, I'll discuss AI communicating to people, in a shared environment, and how we can use highlighting and possible alternatives as a way to combat confabulations (aka hallucinations). In the second, I'll discuss people communicating to AI systems, and how we can leverage language's capability to describe the same behavior at multiple levels of abstraction. Finally, I'll discuss people and AI interacting at the low level of predictive text systems, and how subtle differences in the behavior of the AI system can - or can not - change people's behavior. Throughout, I'll highlight what I think are interesting open questions that I'd love to discuss further with anyone interested!

Speaker

Hal Daumé III holds the prestigious position of Volpi-Cupal Professor of Computer Science and Language Science at the University of Maryland (UMD), where he leads cutting-edge research at the AI Interdisciplinary Institute at Maryland (AIM) and the Institute for Trustworthy AI in Law & Society (TRAILS). His work focuses on advancing machine learning models that are both interactive and ethically sound, seeking ways to develop systems that learn naturally from human interactions while minimizing societal harms.