Skip to content

Artificial Intelligence in Endurance Sports: The Good, The Bad, & The Ugly

photo of a computer chip representing artificial intelligence

Can AI make you a better athlete?


Artificial Intelligence (AI) has made headlines in recent months with Chat GPT.  AI has been around for a long time starting sometime in the 1950’s when scientists started working on machines storing intelligence later known as computers .  In endurance sports, we’ve seen advancements in AI with regards to GPS watches and apps.  Artificial intelligence tells us how long we need to recover between workouts, what workouts to do on which day, and how our current fitness ranks.

In endurance sports, there are both positives and negatives to AI.  We are going to go over the good, the bad, and the downright ugly when it comes to AI.


The Good:

  • A good introduction to following a training plan

Many athletes who first seek out coaching started out following either a stock training plan or an AI generated training plan.  For example, Garmin Coach allows runners to choose from a 5 km, 10 km, or half marathon training plan distance.  The plan is generated considering an athlete’s current weekly mileage, how many times per week they want to run, and days of the week they can run.  These types of programs offer some customization and are also cost effective.


  • Race predictions and pacing plans

These topics are not applicable to athletes who race off road such as trail ultrarunners or X-Terra triathletes.  However, for athletes such as road runners AI has gotten quite good at predicting race times based on training and racing data.  It also  has the ability to create pacing plans for specific races taking elevation into account.  Stryd has a race calculations table as well as a race power calculator that takes the specific race course into account.  The Garmin PacePro feature will create race split times based on your goal time and the race terrain.  This is particularly useful for hilly courses such as the Boston Marathon.


  • Apps that enhance remote based coaching

One obvious downside to remote based coaching is that coaches don’t see the athletes perform their workouts.  The ability to analyse GPS files and see exactly where the athlete did their workout and have data such as pace, distance, elevation, heart rate, cadence, and power has improved remote based coaching.  There are now apps that use AI to analyze running form, one of these being an app called Ochy.  This allows me as a remote based coach to receive good video of athletes running as well as analysis about their running form.


The Bad:

  • AI does not factor in life stressors and the overall complexity of a human being.

Flexibility is key in training and is needed by athletes nearly every week. There’s a reason why the top athletes at endurance events are not following AI generated plans.  They are either self-coached or they have a human coach.  There are so many things that come up from a life perspective in any given week that will impact what the athlete does in training.  The most common frustration I hear from athletes moving from AI generated plans to custom human coaching, is that they didn’t know how to adjust the plan when they missed workouts due to family or work obligations or needed to reduce training load due to injury or illness.  It can get complicated very quickly.


  • Limitations with regards to multisport and off-road athletes

Road runners and road cyclists have more AI options for training plans.  For example, the suggested workout feature in Garmin asks if you are a runner or a cyclist.  For multisport athletes choosing a runner option, the Garmin will always suggest workouts based on the running you are doing and will not factor in the other training load from cycling, swimming, and strength training.  Trail ultrarunners also will not have as many options with AI generated plans as the race distances can be up to 200 miles or more, and the race terrain and elevation become key factors in the training preparation that is needed.  The more complexity your race will have, the less an AI generated plan will serve your needs.


The Ugly:

  • Negative psychological impact of algorithm-based performance and fitness metrics

If I had a dollar for every time I heard an athlete stress about what their watch told them about their VO2 max value, their performance, and their fitness I would be a millionaire.  The last thing an athlete needs to hear when they are tapering for a big race is a beep from their watch stating that they are underperforming and their Vo2 max is declining.  The algorithm doesn’t understand rest is necessary to peak for a race, and reducing training volume is needed during times of injury or illness.  AI treats humans like robots that can increase fitness levels into eternity.  No recovery weeks needed, no job to tend to, or family to take care of.


  • No Human connection

What makes you feel more satisfied – an exchange over social media in your direct messages or meeting up with a friend for a run?  If we’ve learned anything about social media and our time on Zoom throughout the pandemic, it’s that human connections make us feel good.  We need them for not only our mental health but also our survival.  In Brad Stulberg’s book, The Practice of Groundedness he writes, “Our social connections and our sense of belonging – our deep community – influence everything from our physical and mental health to our performance to our life satisfaction and fulfillment.  We evolved to be in community. It is what holds us as we rise and fall.  When we neglect it, we do so at a great cost.”  He goes on further to quote the Zen master Thich Nhat Hanh.  “Without being in a community, without being supported by a group of friends who are motivated by the same ideal and practice, we cannot go far.”


The Future?

Coaching is so much more than a training plan.  Humans are complex, and in turn, so is coaching.  There is a psychological component, social component, and biological component to coaching.  In my opinion, AI only has part of the biological component covered.  I could make all of my coaching income from creating an app or selling courses or training plans.  However, to be honest it wouldn’t be nearly as fulfilling to me as the human connections I’ve made through coaching athletes are irreplaceable.  Looking someone in the eye and saying you believe in them 100%; lending an open ear during hardships such as loss of a loved one, job loss, injury, or illness; adjusting training plans and goals due to life stressors and hardships – these types of human connections have evolved over thousands of years and will not be replaced by AI.  The human connection is why I’m here as a coach.  I have adopted AI with regards to its benefits, but I won’t change the core values of why I started the Performance Project in the first place – having a personal relationship with athletes and supporting them as human beings.

Leave a Reply

    Your Cart
    Your cart is empty