CHI 2026 Best Paper: AI Uses Electrical Muscle Stimulation to Physically Guide Humans

MacrinePhD · x · 2026-07-22

Awarded Best Paper at CHI 2026, researchers demonstrated a novel embodied AI interaction system. Using real-time multimodal AI (body pose, vision, location), the system dynamically generates physical guidance via Electrical Muscle Stimulation (EMS) to assist users with tricky physical tasks.

For example, when mounting a bike on a bus rack, instead of relying on pre-programmed movements, the AI infers the necessary actions and coordinates the user's muscles step-by-step: flexing the biceps to reach, engaging fingers to grip, and firing the triceps to lower the rack. This represents a significant step toward AI taking physical form to assist humans.

Related event: CHI 2026 Best Paper: Embodied AI Guides Human Actions via EMS(3 posts)→

Original post →

More from Embodied

Embodied channel →