I never expected artificial intelligence to help me relearn how to walk.

After an osteotomy on my left second metatarsal, something about my gait felt wrong. I had previously undergone the same surgery on my right foot, yet this recovery was different. I could feel myself compensating on the left side, trying to match my right leg’s rhythm. I wasn’t limping exactly. In fact, I had become remarkably good at disguising the problem. I didn’t know what I was doing biomechanically. I only knew that I was faking a normal walk.

As time passed, my hips and lower back began to hurt. I told my foot surgeon. I later told spine and hip specialists. All of them watched me walk. None could explain why I felt so unstable.

Eventually, my foot surgeon referred me to physical therapy for gait training. My therapist noticed that my left foot pronated and prescribed standard post-operative exercises. They helped somewhat, but they didn’t explain what was actually happening.

Curious and increasingly frustrated, I uploaded videos of myself walking to ChatGPT. Its analysis was startlingly specific. It pointed out that I was unloading my left leg, failing to strike with my heel, landing instead on my midfoot, and not pushing off properly through my toes. More importantly, it explained how those compensations affected the rest of my body.

Over the following months, I continued uploading videos to monitor my progress. Each time, the analysis became a checkpoint. My gait gradually improved. Eventually, walking felt functional again, but something still wasn’t right. The timing between my left and right hips felt subtly off.

Before giving up, I uploaded one final video. This time ChatGPT suggested that I wasn’t “living in” the middle phase of my stride. I was rushing through it. It recommended exercises to help me experience that phase more deliberately. That proved to be the missing piece. Within weeks, walking felt natural again.

Months later, my left knee began hurting. An MRI showed an old MCL injury and cartilage changes. During physical therapy, I realized there was another compensation I had developed: while bending my knee during squats and dance turns, I wasn’t fully loading the muscles that stabilized the joint.

Recovery, it turns out, isn’t one breakthrough. It’s a series of them.

What surprised me most, though, wasn’t that AI had been helpful. It was how differently people reacted to that fact. My friends were impressed. Several female clinicians were curious. They saw someone who had taken an active role in her own recovery.

One physical therapist had a very different response.

“There is nothing wrong with your knee,” he insisted.

When I explained how I had done months of AI-assisted gait analysis, he questioned whether it counted as real gait training at all. Rather than asking what I had learned or whether it had worked, the conversation became about defending professional authority.

The irony is that I wasn’t trying to replace physical therapists. I was trying to solve a problem that traditional care had failed to solve.

This experience made me wonder whether AI threatens more than individual professions. It redistributes expertise. It gives patients access to explanations that once belonged exclusively to specialists. It encourages questions instead of passive acceptance.

For generations, medicine has often expected patients to trust authority. AI changes that relationship. Patients can arrive informed, ask better questions, and participate more actively in their own care.

Perhaps that is what some people find unsettling. Not that AI is replacing experts, but that expertise is becoming more accessible.

I don’t think AI threatens medicine. I think it threatens gatekeeping. For some people, particularly those who equate expertise with personal status, seeing a woman use AI to solve her own problem can feel like a loss of authority. That isn’t a failure of AI. It’s a challenge to the belief that knowledge, and power, should belong only to those at the top of the hierarchy.

AI did not heal me. Exercise, persistence, and countless hours of practice did that. But AI helped me understand what my own body was trying to tell me. That wasn’t a threat to medicine. It was an expansion of my own agency.

Natalia Eden is a nurse.