AI seems to be making news headlines every day. Just this morning, I watched a video of two AI-powered robots putting dishes away with a "shared mind" (sign me up for that!). While we shouldn’t worry about humanoids teaching a BODYPUMP™ class anytime soon, nearly every industry – including our own – won’t be immune to the changes AI will bring.
To find out more, I caught up with Amber Taylor, Chief Digital Product Officer at Les Mills. With over 20 years of experience in international e-commerce at companies like Amazon and Nike, I was eager to hear her thoughts on how AI is shaping the fitness industry.
Jon: How is AI showing up in our industry?
Amber: AI is appearing in fitness in a variety of ways. It ranges from biometrics providing recommendations for what people should do at home or in a class setting, to DNA and blood testing for tailored fitness insights.
Jon: Group training is highly human-centric. Is there a danger that AI could remove some of the human element?
Amber: It’s a nuanced answer… I think we have an exciting opportunity to look at how AI can augment and enhance the human element. If we look at big industrial shifts – like the introduction of cloud computing or mobile phones – each step introduced new fears and new opportunities.
It’s up to us to embrace the opportunity and create what’s right for people. If we’re not curious about what we can do to continuously improve, we miss opportunities. That said, I do not think machines can replace humans, especially in a group training environment – people need people. In fact, I believe this is truer now than ever.
The beauty of group training is that it is group. People crave community and perform better in a community. So that element shouldn’t go away.
But there may be some things we can do to help improve. Things like class counting (a system for automatically counting how many people attend a class leveraging computer vision or other technologies), helping Instructors prep with recommended playlists, eventually form auditing for safety, support with recognition (via things like identifying if someone is on their 10th or 100th class live or digitally and giving a shout-out), creating the right class schedule based on what people attend (leveraging class count) and what’s trending in a given area at specific times of day, and much more.
Jon: Has Les Mills been experimenting with AI, and in what way?
Amber: Yes, in a couple of areas. Internally, the really nerdy stuff is we use AI to help augment our engineering capacity and support with coding. We also use it to help filter our customer service queries for trends so we can fix things faster upstream. We’re exploring AI to help with translations and dubbing of content. For Movers, in the LM+ app, we use AI for rolling recommendations based on explicit intent – what Movers told us in their onboarding survey; and implicit behaviour – what Movers DO.
At Les Mills, we focus on finding the most effective ways to help people start and enjoy fitness. This includes using AI-driven recommendations and gamification, while ensuring safety remains a top priority.
We also tested AI with content. We made a series of modules called “Mindful Movement”, where Movers can do a down regulation breathing exercise with the support of a virtual orb. In testing, we found this reduced heart rate.
Every one of our programs is backed by science. The neat thing is when we do use AI to give recommendations, it’s predicated on safe content that works.
I do not think machines can replace humans, especially in a group training environment – people need people. In fact, I believe this is truer now than ever.
Jon: What have you learnt so far?
Amber: AI is becoming increasingly capable, but it’s still just a tool. For example, we found that translation processes were time-consuming and costly, so we explored AI automation to streamline them. However, AI can make mistakes, which means human intervention is crucial. In one instance, AI mistakenly translated arms as guns in a particular country – not good! That’s why humans need to audit and refine AI outputs.
Another critical lesson is ensuring AI models are trained on diverse and high-quality data. Some companies have faced serious issues due to biased training data. A well-known case (which I won’t name) involved an AI-powered credit card system that was primarily trained on men’s financial data, resulting in women being unfairly denied credit. That’s a major problem.
We can learn from these mistakes by ensuring AI is trained on diverse datasets to produce accurate and fair outcomes.
I think we have an exciting opportunity to look at how AI can augment and enhance the human element.
Jon: Looking to the future, how might AI aid us as Instructors?
Amber: While I can’t share specifics about our AI developments in this space, we are always exploring tools to make our Instructors' lives easier. One potential area is using AI to accelerate and streamline choreography learning. This could include automatically detecting key moves in Les Mills Masterclass videos and breaking down complex sequences into smaller, repeatable chunks for easier learning.
Exciting advancements in image recognition technology could also allow webcams or phone cameras to provide instant feedback on form, tempo, and accuracy during release video playback.
As we discussed earlier, group training is a deeply human profession. Every Instructor is unique and brings their own personal flair to a class. AI must never dilute that. However, if there are opportunities to free Instructors up to focus on what they do best, we will always consider them.
