top of page
Writer's pictureDave Schell

My Take on AI in Endurance Sports: Navigating the Hype and the Hope




As a long-time coach and athlete in the endurance world, I've been keeping a close eye on the rise of artificial intelligence (AI) and its impact on our sport. In a recent podcast, my pal Ryan Cooper and I really dug into this topic, exploring both the exciting possibilities and the crucial limitations of using AI in training.


One thing we've noticed is AI's fondness for analogies to sound more human. While this can make AI-generated content feel more relatable, the comparisons often miss the mark - they're either too vague to be meaningful or too niche to resonate with most folks. It's a reminder that while AI can mimic human speech patterns, it lacks the deeper understanding that comes from real-world experience.


This ties into a larger issue: AI's tendency to present information confidently, even when it's not entirely accurate. In our experiments, Ryan and I have encountered countless examples of AI-generated content that sounds authoritative but doesn't quite nail it. This underscores the vital role of human oversight. As coaches and athletes, we can't simply take AI at its word - we need to bring our own expertise to the table, fact-checking and contextualizing AI outputs before putting them into practice.


It's also crucial to recognize that AI models are only as reliable as the data they're trained on. If that data is biased, incomplete, or inaccurate, the AI's recommendations will reflect those flaws. In the realm of endurance sports, this could lead to training plans that are ineffective at best and harmful at worst.


This is especially relevant when we consider the proliferation of wearable tech and the data it generates. Metrics like wrist-based heart rate and heart rate variability (HRV) have become increasingly popular, but they're not always reliable. I've seen countless examples of athletes getting caught up in the hype, obsessing over every data point without considering the potential for errors.


As coaches, one of our key roles is to review this data with a critical eye. We need to be able to spot outliers and erroneous readings, filtering out the noise to get to the true signal. This applies not just to heart rate data, but to power metrics as well. A single wonky reading can throw off an entire analysis if we're not careful.


The risk, of course, is that if we feed flawed data into an AI system, we'll get flawed recommendations out. An AI trained on a dataset full of inaccurate wrist-based heart rate readings, for example, might suggest training intensities that are way off base. If we're not vigilant, we could end up with a classic case of "garbage in, garbage out."


So how do we navigate this landscape? It starts with a commitment to data quality. Those of us working on AI applications in endurance sports need to be meticulous in our data collection and curation, striving to build datasets that are accurate, unbiased, and representative of the full spectrum of athletes we serve. This means being discerning about the wearables we use and the data we include, always favoring accuracy over mere quantity.


Rigorous testing and validation are also essential. Before deploying an AI model, we need to thoroughly test its performance on fresh data to ensure its recommendations are genuinely generalizable. And once a model is in use, ongoing monitoring is crucial to identify and correct any drift from intended outcomes.


At the end of the day, the responsible use of AI in endurance sports requires a human-centric approach. We can't just feed data into a black box and trust whatever comes out. We need to bring our domain expertise to bear at every stage - from data selection to model development to real-world application.


This is where I see AI's true potential in our sport. When used judiciously, AI can be an incredibly powerful tool to support and enhance human expertise. In my own coaching practice, I've used AI to streamline data analysis, generate course demands, and look for potential limiters. By focusing AI on specific tasks where it can add efficiency and insight, we can free up more time and energy for the high-touch, deeply personal aspects of coaching that no machine can replicate.


Looking ahead, I'm excited about the possibilities. As more coaches and athletes begin experimenting with AI, we'll see the technology continue to mature and specialize. We may soon have AI tools that can optimize race nutrition, predict peak performance windows, or identify injury risk factors on an individual level.


But even as AI becomes more sophisticated, I believe it will always be a complement to, not a substitute for, human expertise. Endurance sports are fundamentally about the human experience - the passion, grit, and resilience it takes to push ourselves beyond perceived limits. AI can support and enhance that experience, but it can't replace the wisdom and empathy of a skilled coach or the self-knowledge of a dedicated athlete.


My advice to those eager to explore AI in their training? Start small, focusing on specific areas where AI can add value. But always maintain a critical eye. Verify AI-generated recommendations against your own knowledge and experience, and never take wearable data at face value without a closer look. Use AI insights as a starting point for further exploration, not as unquestionable truth.


Above all, remember that AI is a tool, not a panacea. Used responsibly, it has immense potential to advance the state of endurance training. But realizing that potential will require an ongoing commitment from all of us - coaches, athletes, and technology developers alike - to keep the human element at the heart of our sport, even as we embrace the power of data and AI.


The future of endurance training is undoubtedly intertwined with the evolution of AI and wearable tech. But it will always be grounded in the timeless human qualities that have defined our sport for generations: dedication, resilience, and the relentless pursuit of personal excellence. As we navigate this new frontier, let us strive to use these tools in ways that honor and elevate those qualities. That's the vision that excites me, and it's one I'm eager to continue exploring in partnership with the endurance community.


111 views0 comments

Comments


bottom of page