
Back pain is the world’s silent epidemic—common, crippling, and chronically under-treated. In the UK alone, hundreds of thousands wait months for basic musculoskeletal care, with the backlog stretching the NHS to its limits. Globally, lower back pain is the leading cause of disability, affecting over 220 million people and draining both economies and personal well-being. The demand for treatment far outweighs the supply.
Now, a new solution is making waves: the AI physio. Tools like Flok Health’s artificial intelligence–powered physiotherapy app promise faster, scalable access to care. Instead of waiting weeks for a referral, patients can begin treatment instantly—straight from their phone. No need for scans, no clinic visits, no hands-on adjustments.
But can an AI physio truly heal a real injury? I tested Flok alongside traditional physical therapy in the U.S.—and what I discovered reveals both the promise and the pitfalls of automated care.
When AI Joins the Clinic: Which Profession Can’t AI Replace?
Back pain isn’t just a nuisance—it’s a global crisis. With over 223 million people affected worldwide, it’s also one of the most common reasons people seek medical help. In the UK, demand is off the charts: more than 350,000 people were waiting for musculoskeletal treatment by late 2024. And that’s where AI physio tech is stepping in—fast.
Flok Health, a UK startup, is now treating patients through a fully AI-enabled physio clinic. It’s the first of its kind approved by the NHS and the Care Quality Commission. Patients use an app where a virtual physiotherapist guides them through stretches and pain assessments—all stitched together through AI.
But here’s the big question: can AI replace physical therapists entirely? While it might streamline care, many experts argue that human empathy, intuition, and real-time correction—hallmarks of hands-on physiotherapy—aren’t so easily replicated. It’s raising new debate about which profession can’t AI replace—and whether physios are safe from the robot takeover.

Behind the Curtain: My Experience with a Virtual Physio
After injuring my back lifting a bed last fall (rookie mistake, I know), I entered a months-long cycle of pain, specialist referrals, and finally, in-person physical therapy. Weekly sessions with a human physio became my salvation: real-time corrections, massage therapy, and tailored exercises helped tame my sciatic pain.
But curiosity (and convenience) led me to try something new: Flok’s AI-powered physio app. Instead of live interaction, I was greeted by Kirsty, a calm, pre-recorded virtual physiotherapist. She asked questions, guided me through stretches, and recommended exercises—all tailored to my answers in real-time.
The catch? It’s a one-way interaction. Kirsty can’t see my form or correct my posture. Feedback is based on what I report. That’s both the promise and the pitfall: it scales care fast but lacks the nuance of physical observation. For tech-savvy users with mild conditions, it’s a sleek solution. But for injury-prone klutzes like me? It’s no substitute for a trained eye.
Stretching the Limits: AI Physio and the Future of Back Pain Treatment
AI physio is no longer sci-fi—it’s happening now. Flok Health’s app is the first AI-backed physiotherapy clinic approved by the NHS, and its goal is clear: tackle the musculoskeletal crisis by treating simpler cases through automated care. The service now treats patients across the UK and plans to expand into hip, knee, and women’s pelvic health.
Flok doesn’t use ChatGPT-style AI. Instead, it relies on a custom logic engine built around clinical reasoning. That means no hallucinations or rogue recommendations—just structured, adaptive video pathways with over a billion potential combinations.
It’s not alone in the field. Apps like selfBACK, currently undergoing trials in the US, use data-driven treatment recommendations based on what worked for others with similar issues. However, adoption has been rocky. Many users skip or abandon the app, revealing a key insight: AI physio works best when paired with human care.
This raises an essential question for those studying AI in physiotherapy courses: Can machines enhance recovery without replacing human touch?

AI and Healthcare: A Balancing Act Between Access and Accuracy
Medical experts remain cautiously optimistic about AI’s role in healthcare. Elizabeth A. Stuart of Johns Hopkins argues AI tools must meet the same gold-standard trials as any other intervention. They can’t just work in theory—they need to work consistently, for real patients, and across demographics.
One challenge? AI tools evolve. Unlike medications or annual vaccines, some AI systems continuously adapt, making it hard to track what’s changed. That can be dangerous in clinical settings where precision matters.
Others, like Harvard’s Pranav Rajpurkar, believe the real win lies in clearly dividing tasks between AI and clinicians. Instead of blending roles awkwardly, let AI handle triage or repeatable routines, freeing up doctors for critical, complex cases.
And that’s where Flok shines. Patients can send voice notes for follow-up from a real physiotherapist. This hybrid model—tech-forward, human-backed—may be the most realistic and ethical future of AI in medicine.