Changing the Game: The Impact of AI in Sports

The use of technology in sports is on the rise and one of the most exciting advancements in this field is the integration of artificial intelligence.

AI is transforming the way athletes train and the way sports organizations run their teams. By analyzing vast amounts of data, AI algorithms can provide insights into an athlete's performance and help coaches make more informed decisions.

Let's take a look at the various ways AI is currently being used in sports and the potential impact it could have on the future of athletics.

Two-time Olympic Triathlon gold medalist and Triathlon World Champion Alistair Brownlee has been exploring how AI tools like ChatGPT can help improve his performance.

To find out, he turned to ChatGPT for training advice;

"Create an Olympic distance Triathlon training program for me. The event is in 12 weeks, I am a 34-year-old experienced male Triathlete. I have 20 hours a week to train."

He was impressed with the fact that ChatGPT was able to present a training program on the screen within a matter of seconds. The program followed a well-organized, progressive, and structured approach that included both easy and challenging exercises. However, he noted that it could be perceived as generic and not tailored to his individual fitness objectives or workout preferences and said that it may be similar to finding an already published training program on Google.

He said that the ChatGPT-produced training program fails to incorporate what he believes are the actual advantages of utilizing AI to prescribe training. He pointed out that most athletes record vast amounts of data such as power, heart rate, and velocity, as well as data from wearables, HRV, resting heart rate, and breathing rate. According to him, AI could successfully analyze this data, compare it to trends, and use the results to suggest adaptations to future training, which is a significant benefit.

That's where a platform like Athletica comes in.

Athletica is a science-based AI coaching platform that's tailored to athletes' specific needs. It takes into account their current sport, fitness level, goals, training sessions, and even their lifestyle and comes up with personalized training plans.

It also allows them to connect their wearable Garmin or Strava devices so that they can collect their training data and continuously adjust the stress of the training sessions from there to ensure the best performance possible.

The athletes can also input feedback after each session so the AI can adjust the level of the workout as well.

The platform is designed either for individual athletes or for coaches.

Jo Clubb is a Sports Science Consultant at Global Performance Insights who has worked in the Premier League, NHL, and NFL, and now she is a consultant. So she does a lot of sports science writing and data.

She demonstrated in her Youtube video how ChatGPT helped her with her sports science research and writing that is meant to inform athletes and coaches.

She also recognizes the importance of monitoring training load for athletes and asked ChatGPT to explain its importance:

"Why monitor training loads in Athletes?"

Which came up with:

"Monitoring training load help to prevent injury and optimize performance, by ensuring that the athlete is not overtraining or undertraining. By tracking the amount and intensity of training, coaches, and trainers can adjust the athlete's training program to ensure that they are getting enough rest and recovery, while also challenging them to improve. Additionally, monitoring training load can also help to identify patterns in the Athlete's performance such as a decline in performance that may indicate the need for additional rest or a change in the training program."

She then went through a series of informative prompts like;

"Discuss the pros and cons of different types of technology to monitor external load. Present this information in a table."

And;

"Can you define acute chronic workload ratio?"

The acute chronic workload ratio (ACWR) is a measurement used to track an athlete's training load and its effect on their performance and risk of injury. It compares the amount of training an athlete has done in the short-term (acute workload) to the amount they've done over a longer period (chronic workload).

So while Clubb is using ChatGPT as her research assistant for information that she later published in Sports Science research papers, the answer for this one actually brings another interesting use case to mind; actually taking an athlete's workout data and coming up with the ACWR number to give an Athlete their risk of injury. So I went to ChatGPT to test it out. I typed in;

"I am a novice runner and I plan to run a 5k in 6 months. My plan is to run 1 mile, every other day for 1 month, and then gradually increase. In my first month, what is my acute chronic workload ratio?"

The response gave me an accurate calculation, along with an explanation, and then also proceeded to give me advice about the final number, so I would say that is a success!

"Your acute workload is the total distance you run in the last week of your training period (every other day for one month). Assuming you run 1 mile every other day for four weeks, your acute workload for the first month would be:

Acute workload (AW) = (1 mile/day) x (7 days/week) x (4 weeks) = 28 miles

Your chronic workload is the average of your total distance run in the past four weeks. Since you only ran 1 mile every other day for one month, your chronic workload for the first month would be:

Chronic workload (CW) = (28 miles) / 4 weeks = 7 miles/week

Therefore, your acute chronic workload ratio (ACWR) for the first month would be:

ACWR = AW / CW = 28 miles / 7 miles/week = 4

An ACWR of 4 suggests a very high training load, which may increase your risk of injury. As a novice runner, it's important to gradually increase your training load to avoid injury and burnout. Consider working with a coach or following a structured training plan to help you gradually increase your workload and improve your fitness."

Following ChatGPT's advice, maybe it would be better for me to work with a platform like Athletica!

If you've seen the movie Moneyball, you'll know that statistics can be used to build a successful team. The movie tells the story of how the Oakland Athletics, a Major League Baseball team, used statistics to build a successful team despite having a limited budget. By focusing on in-game data, the team was able to identify players who were undervalued by the market and assemble a winning team that made it to the playoffs in 2002.

Thanks to advances in technology and the availability of large amounts of data, sports organizations are now able to use artificial intelligence and machine learning algorithms to gain insights into player performance, make better business decisions, and improve overall team performance.

Patrick Lucey, the Chief Scientist at Stats Perform, a machine learning model trained on a large amount of sports data says that while ChatGPT lacks the specific sports data and language needed to perform this kind of sports analysis, Stats Perform is using generative AI and their own proprietary language of sport as the input to generate an analysis of things like player and game performance.

This technology can help coaches predict how well their players will do in a game, and how the team will perform overall by using a large database of sports data that includes both historical and real-time data. This data is fed into machine learning models that analyze patterns and make predictions based on those patterns.

Big names like the Orlando Magic used Stats Perform have used Stats Perform's platform to help them prepare for the NBA draft.

Similar to the case above, AI is also changing the way people watch sports too. AI sports companies like Second Spectrum are using artificial intelligence to provide relevant stats and graphics on screen, in close to real-time.

The NBA, MLS, and Premier League have partnered with Second Spectrum to offer a new, more modern way to watch sports. It comes at a time when the NBA is desperately trying not to lose younger viewers.

In March, the NBA debuted a Second Spectrum AI-assisted broadcast feed on ESPN+, branded Full Court Press Second Spectrum for a game between the Los Angeles Lakers and Milwaukee Bucks. There were three different viewing modes available that each highlighted different information graphically which they say will be an option for the viewer to switch between in the future.

It works by tracking players using cameras that have been installed in every arena, then uses computer vision technology to log the position of each player and the ball at every moment of the game. This data is translated into moving dots which creates the data set that Second Spectrum AI uses machine learning to analyze which is then used for generating the images for the sports broadcast.

Currently, there is a delay of approximately two minutes between the live game feed and the Second Spectrum feed, but the company is trying to reduce that time.

So as we've seen, AI has the potential to be a game-changer in the world of sports. From enhancing athletic performance to providing real-time analysis and predictions, the possibilities are endless.

But let's not forget the importance of the human element. While AI can provide valuable insights and data, it's up to the athletes and coaches to ultimately make the decisions and execute the game plan.