I'm better than AI at rehab. Any day of the week, one hand tied behind my back, hopping.
I also cost a fair chunk more. Reassuringly expensive you might say. Can I quantify that delta for the bean counters who care? Nope. So I'll have to prove it the old-fashioned way.
If rehab really is just a list of exercises, AI wins. Easily. Every time. But it isn't.
Rehab is spotting fear before anyone says it out loud. It's seeing the worry behind someone's eyes. It's watching how they move when they don't know you're watching. It's knowing when to push, when to back off and when to change tack completely because something doesn't add up. Touch, eyes, ears, environment, trust, you get my drift. It's getting someone to actually do the rehab, not simply delivering it to them.
So AI isn't replacing physios. It's replacing exactly the version of physiotherapy we've spent decades building.
We told everyone the exercises were the treatment. We told them the hands-on bit had no effect because the research said it didn't change the tissue. It got taught out at university and designed out of NHS pathways. Patients never agreed with us. They kept asking for hands-on and we kept explaining why they were wrong.
The issue is manual therapy was never the mechanism. It was the vehicle. It bought attention, permission, trust and adherence, and it made people believe the rest of it was worth doing. We disproved the engine and threw away the car. What we're left with is a list of exercises and a rationale, which is a beautifully machine-readable product.
And that's what interests me. Every physio I seem to meet is excited about AI. Good, so am I. But when an entire profession agrees on something, my spidey-sense starts to tingle.
For mass MSK waiting lists I completely get it. AI can give thousands of people something when right now a lot of them are getting the square root of bugger all. That's progress. Private practice is different. If someone is choosing to spend £100, £150 or £250 of their own money to see us, we had better be offering something fundamentally different from a clever algorithm and a beautifully delivered exercise programme. Because if we're not, I wouldn't pay us either.
Perhaps the threat isn't that AI gets better at physiotherapy. Perhaps it's that we've spent the last twenty years making physiotherapy easier for AI to do.