AI-Assisted Physiotherapy Showed Mixed Results in Trial

A study compared AI-designed shoulder pain exercises against conventional methods with varying outcomes.

Updated on Sept. 26, 2026 in Fitness

Resistance bands and a medicine ball in a clean, bright, modern physical therapy rehabilitation studio.
A university-led trial found that AI-designed exercise programs for shoulder pain yielded outcomes comparable to conventional physiotherapy methods over a six-week period. AI Illustration. Upload story photo >

Live Poll

Do you trust AI-generated programs to effectively manage your physical pain as well as a human?

Researchers found no significant difference in overall shoulder pain and disability scores between AI-assisted and conventional physiotherapy programs. However, AI-driven programming did lead to improvements in resting pain levels.

Why it matters

The trial sought to determine if AI tools like ChatGPT-4o or DeepSeek-R1 could match or outperform standard human-designed exercise plans for subacromial pain syndrome. The findings suggest AI can play a role in pain management but may not universally replace traditional approaches.

The study involved 75 participants with 69 completing the 6-week follow-up. While overall disability scores showed no significant difference (P = 0.428), DeepSeek-R1 programming outperformed conventional methods in resting pain reduction (P = 0.005).

The players

ChatGPT-4o

This is a large language model developed by OpenAI that was used to assist in designing the exercise programs for the trial participants.

DeepSeek-R1

This is an artificial intelligence model that outperformed conventional programming in reducing resting pain levels for patients during the study.

The details

Participants attended supervised exercise sessions three times per week for six weeks at a university-affiliated hospital. The researchers measured changes in the Shoulder Pain and Disability Index to compare clinician-curated AI programs with traditional therapy.

Timeline

  1. August 3, 2025: The clinical trial was first posted to ClinicalTrials.gov.

  2. September 26, 2026: The research article was published online.

The Big Picture

This study follows established protocols like the Shoulder Pain and Disability Index to quantify how digital tools are integrating into clinical physical therapy. It highlights an early transition point where AI is being tested for therapeutic efficacy rather than just administrative task support.

Patients may eventually see AI tools used to customize their recovery exercises, potentially offering more effective relief for specific symptoms like resting pain. However, these tools are currently used as assistants by professionals, meaning patients should still consult a licensed therapist.

The takeaway

AI tools show measurable promise in specific pain management areas, suggesting a future where digital assistance supplements physical therapy. Patients should remain aware that AI currently serves as a tool to aid clinicians rather than a standalone replacement for professional guidance.

Further reading

For more information on the intersection of technology and exercise, visit the Fitness section.

More information

Read the full results in the peer-reviewed research article.

Source note: This article includes information reported by Nature.

Live Poll

Do you trust AI-generated programs to effectively manage your physical pain as well as a human?

AI-Assisted Physiotherapy Showed Mixed Results in Trial | Wisevoter