The AI Debate We Need to Have in HIGH PERFORMANCE SportS

High Performance SportSports ScienceTechnology in SportCoaching PhilosophyTools & Systems
Written By Jason Weber

Strong Thinking First. Tools Second.

In elite sport, fads come and go. I’ve seen them all over the past 30 years…from vibrating platforms and cryo chambers to tech tools that promised insights and delivered noise.

So when AI in sport started making headlines, I didn’t jump in. I sat back... not to write it off, but to observe. And now that it's hitting the mainstream in high-performance environments, I think we need a straight-shooting conversation:

AI is here... but critical thinking is still king.

It’s not going to replace the human element in coaching, sports science, or rehab. It can’t read context. It doesn’t know your athlete. And it definitely can’t replace professional judgment.

1. AI Writing Your Strength Programs? No Chance

Let’s start with one of the most talked-about use cases… AI auto-generating strength and conditioning programs.

To be clear... I’m absolutely against it.

Programming is not a plug-and-play task. It’s a coaching process built from movement analysis, training age, injury history, session intent, psychological readiness, and more. Reducing all that to a single prompt misses the entire point of coaching.

We’ve seen this in other sports too especially baseball and AFL where load management tools were adopted before understanding what that load meant. The result? Misinterpretation. Over-reliance. And in many cases... more injuries, not fewer.

Could AI assist in customising a program based on inputs you set? Maybe... one day. But right now? We’re not there. And the risk of overestimating what AI can do is high… especially for young practitioners still building their programming instincts.

So here’s the line:

AI can support good programming... but it can’t replace a good coach.

2. "I’ll Just Ask ChatGPT to Code It" — That’s a Red Flag

Here’s the second trap:

“I don’t need to learn how to code... I’ll just ask ChatGPT to do it.”

That mindset’s not just lazy... it’s dangerous.

Yes, I use AI tools like ChatGPT too. But I always start with my own logic. I write the code. I test it. I look for flaws. Then — and only then — I might ask AI to clean it up or check for gaps.

Because when you blindly hand over a problem you don’t understand... you lose control. You’re not working with AI. You’re guessing with it. In high-performance sport, that’s not good enough.

Take data dashboards for example. If you ask AI to build one from scratch using your athlete performance data, it might give you something that looks polished. But under the hood? I've seen it — dodgy formulas, mismatched timeframes, misunderstood variables.

So what do I tell coaches? Do the dirty work first. That’s where the real learning lives. Learn to write your own code... even if it’s messy. That’s what I call guerrilla coding. Build the rough version yourself. Then refine. Then automate.

3. AI Should Amplify Your Thinking — Not Replace It

Let me be clear. I’m not anti-AI. I’m anti-outsourcing your thinking.

In sports science, rehab, S&C, and biomechanics, your biggest weapon isn’t a tool — it’s your ability to think.

To weigh up trade-offs. To choose the right metrics. To ask the right questions when an athlete’s not adapting or progressing.

AI can assist... speed things up... even help you catch things you might miss. But it can’t create clarity. That still comes from you.

This same conversation is happening in track & field with biomechanics assessments. Coaches are being sold automated gait reports, but if they don’t know how to interpret what they’re looking at... it’s just data noise.

To Young Coaches and Sports Scientists: Learn to Code

If you're serious about a career in elite sport, especially in the pro or international setting, here’s the honest truth:

You need to understand your AI tools... not just use them.

Learn to code. You don’t need to become a software engineer. But you do need to know how to work with performance data at scale.

  • Wrangle complex datasets from GPS, force plates, and IMUs

  • Build models that reflect how multiple inputs interact over time

  • Run your own logic through R or Python and know what you’re testing

  • Understand how tools like SpeedSig or ForceDecks output data — and what’s behind the curtain

Basic stats like t-tests and stepwise regressions won’t cut it anymore. We’re dealing with non-linear systems. Machine learning, multivariate analysis, probabilistic models. That’s the landscape now. And AI can support this but it can’t replace your understanding of what the system is doing.


Where AI belogs in Sport (and where it doesn’t)

So Where Does AI Actually Fit in Sport?

Here’s the filter I use when deciding whether to use AI on a job:

✅ Smart Use Cases

  • Optimising your existing code

  • Translating logic into pseudo-code

  • Speeding up reporting and visualisations

  • Automating repeatable tasks (like converting raw GPS outputs)

  • Prototyping ideas to test with your own logic

🚫 Not-So-Smart Use Cases

  • Writing S&C programs

  • Replacing your ability to troubleshoot code

  • Solving statistical problems you don’t understand

  • Making critical training or rehab decisions

  • Using it to impress others instead of educate yourself

AI is evolving. It’s getting better. And in the next 5–10 years, it’ll be part of every high-performance sport environment — from football and rugby to cycling and baseball.

But don’t get it twisted.

AI should live on top of good thinking... not instead of it.

If you’re in the game to stay, sharpen your tools... but sharpen your mind first. Learn how to build systems. Learn how to interpret data. Learn how to lead decision-making under pressure.

AI can move fast. But the best coaches still move smart. — Jason Weber

Jason Weber
Next

Are We Testing or Monitoring?