If you've had a sports injury in the last year or two, there's a good chance you typed your symptoms into an AI chatbot before you saw anyone. You probably also have a watch or app that has logged every run, ride and night's sleep for months.
Both can be useful. But they're very different when it comes to your privacy, and to how much they can actually tell you about an injury. Here is a plain-language look at where AI fits in sports medicine today.
A chatbot is not your doctor's consulting room
When you speak to a doctor, what you say is confidential. Doctors are bound by professional confidentiality rules, and your health information is protected by law (in South Africa, the Protection of Personal Information Act, POPIA).
An AI chatbot is not bound by medical confidentiality. Depending on the service and your settings, what you type may be stored, reviewed by the company's staff or used to improve its systems, and it may be processed outside South Africa. Read the privacy settings of any service you use, and assume what you type may not stay private.
Before you type anything:
- Leave out anything that identifies you: your name, ID number, date of birth, medical aid details or address.
- Don't upload scan reports, referral letters or test results that have your details on them.
- Ask in general terms. "What causes pain at the front of the knee after running?" works just as well as a full personal history.
What a chatbot is good for
- Understanding words in a report you've already been given
- Preparing questions to bring to your appointment
- Learning about a condition you've already been diagnosed with
What it can't do
- It can't examine you. Many sports injuries are diagnosed by how a joint moves, where it's tender and which tests provoke your pain.
- It can sound confident and still be wrong. Answers aren't always up to date or suited to your situation.
Don't let a chatbot talk you out of urgent care. Go to an emergency unit if, after a head knock, symptoms are getting worse, or if you have chest pain or feel faint during exercise, can't put any weight on an injured leg, or have new numbness or weakness in an arm or leg.
Your watch data: bring it with you
Your watch, running app or cycling computer uses algorithms to estimate training load, recovery and readiness. The individual numbers aren't exact, but the history is genuinely useful in a consultation, because it shows what happened before the pain started:
- A jump in distance or intensity. A sudden increase in weekly load is a common trigger for overuse injuries, and your log shows exactly when it happened.
- Changes in terrain or pace. More hills, more trail or a new speed session.
- Sleep and resting heart rate trends around the time you got injured or unwell.
- How you're coming back. Once you're recovering, the same data helps plan a gradual return.
You don't need to send it before your appointment. Simply have the app open on your phone during the consultation so we can look at the weeks leading up to the injury together. You stay in control of what you share.
Where else AI fits in
Reading scans
Software can help flag possible fractures and other findings on X-rays and some other scans. It works as a second pair of eyes: the image is still reported by a radiologist and interpreted by your doctor alongside your history and examination.
Elite sport
In elite rugby, instrumented mouthguards now measure head impacts during matches, and video review helps identify players who need a head injury assessment. Some of this filters down to school and club sport over time.
Predicting injuries: not there yet
There is a lot of research into using AI and machine learning to predict who will get injured. So far, no model can reliably tell an individual athlete that they are about to get injured. Be wary of any app or product that promises it can.
The bottom line
AI can help you understand your injury and your training, but a diagnosis still comes from a proper history, a hands-on examination and the right investigations when they're needed. Keep your personal health details out of chatbots, and bring your training data to your appointment.
If you have an injury that isn't settling, our patient guides explain common conditions and when to get checked.