Interfaces for Embodied AI

Synopsis: In this class you will build real prototypes that allow AI systems to interface with the physical world via the user’s body—as such, this class requires substantial knowledge of interactive systems programming and hardware experience (e.g., microcontrollers, actuators, sensors, etc.).

This is not a traditional “paper reading seminar,” in that most of our emphasis will be on completing hands-on projects, which we complement with paper readings and discussions. The goal is to understand that as AI moves beyond screens and into the physical world, new interfaces are needed to connect machine perception and reasoning with human bodies, objects, and environments.

Thus, you will explore embodied AI from an HCI perspective: how interactive systems can sense physical context, reason about it, and—more importantly—translate computational outputs into meaningful physical actions. Drawing on wearable interfaces, robotics, haptics, shape-changing systems, physiological sensing and actuation (e.g., muscle stimulation and robotic force-feedback), you will explore the complete loop between sensing, AI reasoning, and actuation.