MIT and Purdue researchers build generative AI robotic rehab system to

MIT and Purdue researchers build generative AI robotic rehab system to help 5 million annual stroke patients amid therapist shortage

Strokes leave 5 million people worldwide with long-term motor impairments each year, while a growing shortage of physical therapists limits recovery care. To bridge the gap, Purdue professor Johannes Lachner and former MIT researcher Noah Geiger built a generative AI robotic rehab system.

Highlighted in MIT Technology Review (MIT Alumni News 77), the robot combines generative AI with real-time force feedback to learn directly from human physical therapists. It then delivers calibrated assistance during 3 core exercises: reaching, lifting arms, and daily living tasks.

Lachner emphasizes that the goal is to extend therapists' reach rather than replace them. After successful initial robotic experiments, the team's next phase will build therapist-specific AI profiles and evaluate them in a clinical study with patients who previously received manual therapy.