Is AI Injuring Runners?
As AI coaching apps surge, runners face a hidden side effect: overtraining without accountability
A $10 Billion Industry By 2029
Here’s some crazy numbers to kick things off. Fitness app downloads rose to 25.15 million in January 2025, up from roughly 17 million in 2021 (Sensor Tower Q3 2025).
AI-driven fitness coaching apps are projected to dominate the digital fitness market by 2027, with the global app economy driven by real-time feedback and personalisation (Research Intelo 2025).
A $10.06 billion industry by 2029
It’s clear this rise won’t stop anytime soon, with the revenue of fitness apps anticipated to grow at a compound annual growth rate (CAGR) of 7.96% from 2024 to 2029, leading to an estimated market size of $10.06 billion by 2029 (3DLOOK AI Fitness Report).
You could almost say the industry is overloading with this rapid growth in AI-Powered training apps made accessible to the everyday runner.
And what happens when we overload at an excessive rate? We get injured.
2025 Was the Year I Learned This Lesson Myself.
This year I was setting myself up to push my limits beyond what I achieved personally in 2024. Coming into the new year I was primed with consistent mileage and ready to begin prep for my next marathon PB - sub-3:30.
At the time I was a big fan of the app Runna - the running app “built by real coaches”, and also powered by AI. I used the app in 2024 to build my training plan for the Edinburgh marathon. I followed approximately 70-75% of the recommended runs from the plan, trying to stick to my own run club’s weekly sessions in a hybrid training plan approach.
Come race day, I hit my new PB of 3:34 and couldn’t have been happier. So, taking this success to my training in 2025 and pushing the plan to increase my mileage seemed to only be logical.
This is where I went wrong.
The new plan included additional runs to grow my weekly mileage, and much higher intensity speed workouts from the very start. 6 weeks into the plan I noticed my legs were beginning to ache more so than normal. Fast forward a week or so later, I got out of bed one morning and agony hit my right leg. Walking was incredibly painful and I had no form of strength in that leg to propel myself into any form of running motion. I was injured. Properly injured.
After a period of rest I gave in and booked a private physio appointment, to receive an early diagnosis of medial tibial stress response in my right leg. I was out of the running game for the next 4 months, my training plan derailed and my marathon PB went out the window.
Through consultation and reflection of my training up until the injury, it was very apparent my body was being overloaded and pushed well beyond its capacity with my training plan. So I looked into research on training apps and one finding was clear, even without research:
Apps like Runna, Nike Training Club, and Garmin Coach do not currently provide true real‑time workload management or fatigue prediction (Sports Techie or wearable analytics journal review 2024–2025).
The Broader Landscape of Running Injuries and AI
As of 2025, consistent global findings show that overuse, poor biomechanical form, and inadequate recovery remain the dominant causes of running injuries among everyday runners. Roughly 40-50% of recreational runners report at least one running-related injury each year (Predictors of Running-Related Injury, 2024; JOSPT, 2021).
A review of the app Runna from Runner’s World highlighted that some runners have deleted the app after perceiving that training loads increased too quickly, leading to fatigue or injury. While others appreciated Runna’s algorithmic adjustments, physiotherapists have emphasised that the app is still reactive, not predictive, so fatigued runners often train beyond safe limits… leading to injury.
My biggest takeaway from this experience however was that ultimately, I placed total trust in my training app. Accountability is important in moments like this, and I take full accountability for my part in this experience. I chose to subscribe to Runna, set up my training plan, and follow it implicitly.
The app’s accountability? Non-existent. It’s an AI-Powered app which doesn’t carry empathy for its user. It markets itself as an app built by “real coaches”. But were those coaches present when I couldn’t walk without pain? Evidently not. And that’s where I’ve since lost trust in the app.
Trust is crucial to runner-coach and consumer-brand relationships.
A report by F. Wachholz from the Journal of Digital Health Research reported that in a 1,200-participant cross-national study, 54% used structured training plans; 25% relied on AI-generated ones. Trust was high for routine tracking but low for adaptive interventions. Perceived “loss of accountability” was a main rejection driver (Journal of Digital Health Research, 2025).
The growing frustration from avid runners who have experienced injury through overtraining as a result of following AI-powered training plans doesn’t stem solely from the injury itself, but the lack of accountability these platforms take.
The opportunity for a remedy here is straightforward - these apps need to implement greater accountability into their marketing and in-app messaging. The reality is runners are experiencing overuse injuries through lack of adaptive coaching or injury management from these apps, which is going to only negatively impact their brand in the long run.
This messaging can come in the form of reminders to check how your body is feeling during the block. Provide opportunities for input on any new aches or pains. Real coaches listen to their athletes and adapt plans when needed to keep them fit and avoiding injury.
This is the time for these apps to begin to implement greater accountability to regain trust from runners like myself who will probably not use the app again anytime soon.
That being said, time to put on my marketing cap as we highlight some findings on the correlation between injuries, trust and fitness apps.
User Accountability & Brand Credibility in Fitness Apps
A 2023 study by Martín et al. in Frontiers in Public Health found that credibility is the most influential factor affecting user loyalty, satisfaction, and willingness to recommend fitness apps.
When users perceived an app as unreliable or irresponsibly designed - causing injury or stress - their trust and recommendation intention dropped dramatically.
Credibility and quality of information ranked above all other factors, including ease of use, aesthetics, and entertainment. This directly links accountability with brand protection (Frontiers in Public Health, 2023).
Psychological Accountability and Brand Engagement Study: Working (out) with Fitness Influencers and Brand Trust (Durau et al., 2024)
When brands use credible, accountable representatives - like trainers or influencers - user trust in the associated app increases through perceived social accountability.
However, when sponsored influencers cause visible injuries or spread misinformation without accountability, both their credibility and the brand's reputation suffer - demonstrating the reputational cost of poor oversight (Frontiers in Psychology, 2024).
Loyalty & Accountability-Driven Retention Study: Optimizing Fitness App Features for Enhanced User Loyalty (Frontiers in Marketing Psychology, 2025)
This study found that apps with transparent progress feedback, safety warnings, and injury-reporting features retain 30–35% more users than those without accountability measures (Frontiers in Marketing Psychology, 2025).
The risk that AI-Powered training apps begin to run as we see this rise in injuries from runners pushed beyond their limits is increasingly evident.
More Running, More Online Content, More Injury Stories
Jumping back to my own experience, this year I also started a running account on Instagram and exposed myself to a lot more running content online.
I noticed a few posts appearing every now and then by a runner or physio discussing the increasing volume of injuries that runners are reporting while using training apps to guide their training.
There’s even an entire reddit thread dedicated to this topic, asking “How many Runnas are getting injured?”
It was clear, I wasn’t alone in my experience.
But it also needs to be considered that the relationship between AI and injuries currently is deemed a correlation rather than a causation. We’re yet to see a study that accurately links the overtraining through these apps to significant injuries.
The reality is runners get injured all the time, regardless of what app they use. Those stats we looked at earlier on most common injuries were not solely caused by AI and could well have come through human coaches in some cases.
However, that does not excuse the growing rumblings online around how AI-Apps have resulted in runners picking up injuries. And I can’t excuse it myself as one of those runners.
The flip side of AI’s role in running injuries does hold some encouraging signs, but with a caveat at this stage.
A July 2025 study in the Journal of Sports Sciences found that AI can accurately estimate stress loads in knees, shins, and Achilles tendons within a 1.95-12.8% error margin - outperforming manual human estimation in precision.
The caveat? These results came from elite setups using expensive motion-capture technology unavailable to most consumers.
For everyday runners relying on AI apps, injury misclassification rates rise sharply due to missing contextual cues like fatigue and hormonal changes.
The takeaway for everyday runners looking for training with lower risk of injury but are not at the elite level yet, may be to look for your first in-person coach. It has been proven that human coaches remain better at preventing chronic overuse injuries in runners (Journal of Sports Sciences, 2025). While this is taken from one study, it’s important to acknowledge that there has always been a present solution to this rise in AI-related injuries - human coaches.
Closing Thoughts
Bringing all this research together it’s pretty evident when runners follow AI-generated plans without human moderation, they show higher overtraining and chronic strain risks.
Injuries in the pursuit of running greatness are inevitable. The opportunity to prevent them is undeniable.
As someone who recently pivoted into a career in AI, its potential to service runners is an exciting arena of exploration.
AI handles pattern recognition effectively in many contexts. And running is very much one of those contexts.
Within running, given the right inputs, it could handle quantitative load optimisation - a key feature in the story of injury prevention.
Where AI falls short currently, is where human coaches thrive.
Human coaches provide individualisation, empathy, and adaptive approaches to training based on the current state of their athlete. But this is a topic for another post.
Am I saying delete your running app? Not quite. There is certainly a place for them still amongst injury experiences that have been reported. I attribute the Runna app a lot to my early running progression and success in races. And it certainly is a great starting point for runners looking to create their own structured training when a coach isn’t accessible. But there does seem to be a point where we have to weigh up where do we draw the line in trusting its training plans? Especially when those new aches and pains start to creep in.
AI doesn’t cause all running injuries. But it is increasingly present in the lives of everyday runners and therefore cannot be blameless when we see increasing reports of people turning away from training apps after injury or when dedicated subscribers have to start paying for physio.
So where does the opportunity lie for AI in injury prevention with training apps like Runna?
Stay subscribed as we dive into this in part two.