AI-driven robotic therapy aims to extend stroke rehab amid therapist shortage

MIT-developed system learns from human therapists to deliver just the right assistance during rehabilitation exercises

By LineZotpaper

Published20 October 2026

Read Time2 min

Researchers have developed a robotic rehabilitation system that combines generative AI with real-time force feedback to learn how human therapists assist patients, a response to a global shortage of physical therapists that leaves many of the five million people impaired by stroke each year unable to get the care they need.

The system was developed by Johannes Lachner, a former MIT postdoc who is now an assistant professor at Purdue University, and Noah Geiger, a former visiting student at MIT. It uses generative AI and real-time force feedback to learn from real therapists, then gives patients just the right amount of assistance during exercises such as reaching, lifting their arms and performing tasks of daily living.

Lachner said the goal is "not to replace therapists, but to extend their reach."

"Initial robotic experiments have already demonstrated the feasibility of this approach," he reported. "The next phase of the project aims to develop therapist-specific models and evaluate them in a long-term clinical study with the same patients who previously received manual therapy."

Beyond rehabilitation, he envisions possible applications in contact-rich industrial settings or collaborative workspaces.

The work sits at the intersection of healthcare, robotics and artificial intelligence. Strokes are a leading cause of long-term disability worldwide, and the demand for physical rehabilitation consistently outpaces the supply of trained therapists, particularly in underserved regions.

Analysis

Why This Matters

Five million people are left with long-term impairments by strokes each year, and a growing shortage of physical therapists means many cannot access rehabilitation.
The system is designed to supplement rather than replace human care, potentially expanding access while preserving a role for therapists.
If the next phase is successful, the same technology could move beyond rehab into industrial and workplace settings.

Background

Physical rehabilitation after a stroke relies on repeated, guided exercises, often needing one-on-one attention from a trained therapist. The approach being trialled here trains a robot by observing how human therapists provide assistance, then uses that model to deliver support in real time. It is one of several efforts to apply artificial intelligence to healthcare, though its focus on learning directly from clinician behaviour distinguishes it from purely scripted therapy robots. The researchers have moved from early demonstrations to planning a longer clinical evaluation with the same patients who previously received manual therapy.

Key Perspectives

The researchers: Lachner and Geiger frame the technology as an extension of therapists' reach, not a replacement. They point to feasibility demonstrated in early experiments and are planning a long-term clinical study with patients who earlier had manual therapy. Medical and clinical observers: The likely test is whether the system performs as well over time and across different patients as hands-on therapy, and whether therapist-specific models can be built and evaluated rigorously. Skeptics: The approach has demonstrated feasibility in initial robotic experiments but has not yet shown long-term clinical outcomes. Its value will rest on results from the planned study with previously treated patients.

What to Watch

Results of the planned long-term clinical study involving the same patients who previously received manual therapy.
Whether therapist-specific AI models perform consistently across different patients and recovery stages.
Signals of expansion beyond rehabilitation into contact-rich industrial settings or collaborative workspaces.

Sources

Robotic therapy for stroke patients · MIT Technology Review

Checked

12 claims in this story checked against 1 source: 7 supported.

Show each claim

SupportedAI-driven robotic therapy aims to extend stroke rehabilitation amid a therapist shortage.MIT Technology Review: “A growing shortage of physical therapists makes it hard for them to get the care they need. ... The goal is “not to replace therapists, but to extend their reach,” Lachner says.”
Partly supportedThe system was developed by MIT.MIT Technology Review: “Johannes Lachner, a former MIT postdoc who is now an assistant professor at Purdue University, and Noah Geiger, a former visiting student at MIT, developed a robotic system”
Partly supportedA global shortage of physical therapists is a response to which the system was developed.MIT Technology Review: “Strokes leave 5 million people with long-term impairments each year, and a growing shortage of physical therapists makes it hard for them to get the care they need. A new approach tries to solve that problem”
SupportedStroke leaves five million people with long-term impairments each year.MIT Technology Review: “Strokes leave 5 million people with long-term impairments each year”
Partly supportedThe shortage leaves many of those five million people unable to get the care they need.MIT Technology Review: “a growing shortage of physical therapists makes it hard for them to get the care they need”
SupportedThe system uses generative AI and real-time force feedback to learn from therapists and provide patients the right amount of assistance during rehabilitation exercises.MIT Technology Review: “a robotic system that uses generative AI and real-time force feedback to learn from real therapists. Drawing on what it’s learned, it gives patients just the right amount of assistance during physical rehabilitation exercises”
SupportedJohannes Lachner is a former MIT postdoc and current Purdue assistant professor, and Noah Geiger is a former MIT visiting student.MIT Technology Review: “Johannes Lachner, a former MIT postdoc who is now an assistant professor at Purdue University, and Noah Geiger, a former visiting student at MIT”
SupportedLachner said the goal is “not to replace therapists, but to extend their reach.”MIT Technology Review: “The goal is “not to replace therapists, but to extend their reach,” Lachner says.”
SupportedLachner reported that initial robotic experiments had demonstrated the approach’s feasibility.MIT Technology Review: ““Initial robotic experiments have already demonstrated the feasibility of this approach,” he reports.”
SupportedThe next phase aims to develop therapist-specific models and evaluate them in a long-term clinical study with the same patients who previously received manual therapy.MIT Technology Review: ““The next phase of the project aims to develop therapist-specific models and evaluate them in a long-term clinical study with the same patients who previously received manual therapy.””
Not in the sourcesStrokes are a leading cause of long-term disability worldwide.
Not in the sourcesDemand for physical rehabilitation consistently outpaces the supply of trained therapists, particularly in underserved regions.