Use of AI in Strength Training

The conversations surrounding AI in sport and coaching is everywhere and it is only getting louder. Scroll on LinkedIn and most captions and posts are written entirely by AI and it makes me wonder - how much is AI going to revolutionise the way we develop athletes, or is it becoming too much of a shortcut that will hollow out the profession?

Where Does AI Have a Place?

There are genuinely useful applications and roles that AI can play in the coaching space and it's worth acknowledging them properly rather than dismissing them.

For coaches who are time poor, which in reality is most of us... Being able to synthesise research quickly and get across a body of literature faster and identify what is relevant to your practice is extremely valuable, especially for the student coaches who are juggling gaining experience alongside further education or full time work. It can help structure written work, session notes, evaluations and reflective logs, the list is really endless. Testing a programming idea before rolling it out to challenge your own ideas and using it as a tool to find holes in your reasoning before an athlete ever sees the programme are all legitimate uses and they support coaches thinking abilities without replacing it.

That distinction is incredibly important and matters more than it might seem.

The Flatness Problem

But here is where it is worth challenging. Programmes are appearing, ones that are being used with athletes and more often than not they are being used in situations where it could matter a great deal to that individuals development, confidence, and their long term relationship with their sport or training. These programmes have clearly had a lot of AI influence in building them, and on the surface they hold up, the structure is clear and the terminology is correct and it looks to be periodised perfectly for the athletes needs.

But when you look closer at these programmes, there is a flatness to them... A one size fits most quality that shouldn't be there when the whole point of an individualised programme is that is isn't meant to fit the masses, it is meant for the individual.

Which raises a question worth sitting with - what does individual actually mean in individual programming?

Athletes aren't profiles, things can be complex and when they have weeks where nothing seems to be going right outside of training and these can be things that coaches are completely unaware of, unless they've built some kind of relationship where an athlete feels comfortable enough to share them. Niggles, confidence, psychological baggage that they might be carrying. The information that can shape a good programming decisions is often not written down anywhere and lives in that relationship between the coach and the athlete, which is built over time through conversations and human observations.

Language models cannot replicate any of that, they can only respond to the information that is within the prompt they are given and no prompt can capture a person fully.

What does this mean for developing coaches?

Coaching should be a craft, developed through repetition, mistakes, wins, losses and by figuring out why something did or didn't work. This uncomfortable process a lot of us go through is what makes good coaches and forms our instinct and that can not be fast tracked, regardless of how good these tools become. It is where judgement develops and coaches stop applying frameworks and develop the ability to read athletes.

If that process get shortcuts by AI, even partially. What happens to the quality of the coaching coming in the near future? Is it breeding better coaches? or is it moulding coaches to fit the templates that AI is providing and slowly narrowing what coaching should look and feel like as a result?

The Accountability Gap

Having AI within the coaching space leaves a gap which I feel is important and it often gets overlooked.

When something doesn't go right with your programming or an outcome isn't achieved with an athlete, a coach needs to be able to trace their reasoning back, and understand where the thinking broke down and improve for the next block because that reflection is where growth happens, not just for the athlete or the programme but the coach. The feedback loop that can actually improve coaches and make them better over time depends on owning their decisions and reflecting on them personally.

Using a tool to write programmes in their entirety can make that loop murky. The learning gets diluted and for the athlete on the other end they are not receiving an improved version of their coach. They are getting a the same prompt, just slightly adjusted.

So where does that leave us?

None of this means that AI has no place in coaching. It probably means it has quite a significant one but only if it used well.

I believe this requires coaches to know and understand what coaching demands beyond the programme itself and where their human touch is irreplaceable. With the current relationship with AI it feels like this conversation is getting over looked and instead we are debating which tool is best, which is a bit like arguing over which pen to use before you have decided on what you actually want to write.

The hundreds of small decisions, relationships, judgement, presence are all things that dont live inside a prompt. They live in the coach and their experiences.