Artificial intelligence (AI) is seeping into all aspects of our lives, and sport is no exception. This has raised concerns about a loss of authenticity and unpredictability in sporting competitions that are fundamentally human in nature, as well as further widening the gap between well-resourced athletes and the talented underdog. But how valid are these concerns? Let’s take a look at the current field of play.
What do you love most about sport? For most people, the answer lies in human characteristics and emotions. For example, we feel inspired by athletes who dedicate their lives to becoming the best, reaching seemingly impossible levels of skill and physical performance. We revel in a sense of tribalism and belonging when our beloved team emerges victorious against a local rival. And we identify with the plucky underdog who demonstrates courage, hard work, and raw talent to overcome the odds and defeat a more illustrious opponent.
I mean, who would bother watching Rocky IV if Ivan Drago, the human machine ruthlessly trained by analytics and science, had just turned up and smashed Rocky Balboa’s old-school methods of blood, sweat, and tears? It would be pretty pointless because there’s no fun, no drama, and no story in that.
Clearly, sport is most enjoyable when it’s unpredictable, guided by relatable and unquantifiable human traits rather than cold, hard numbers.
The global sports industry
Sadly, this view has begun to appear as hopelessly romantic – naive even – as professional sport has grown into a multi-billion-dollar industry. Huge amounts of money are poured into broadcasting and licensing rights, sponsorship deals, sports betting, clothing and sporting goods, and other sub-markets. The rewards for success are more lucrative than ever, and those with the biggest budgets invariably tend to come out on top.
The rise of data analytics in sports, widely popularized in the early 2000s with the story of the Oakland Athletics baseball team and the subsequent “Moneyball” book and film, has seen many coaches and athletes take a more quantitative approach. Both on-field and off-field analytics are now an integral part of any reputable sporting organization, with research teams dedicated to data collection and analysis of in-game player statistics, physical performance and fitness, fan engagement, ticket sales, and more.
Some argue that this focus on data science is destroying the naturalness and unpredictability that attracts us to sports in the first place. However, athletes are still human (for now) and are prone to making mistakes in high-pressure situations. Data analytics has its limitations on the playing field – but what about when combined with AI?
AI: the game changer?
Artificial intelligence (AI) has reached a tipping point recently, with the technology now sufficiently matured to begin making an impact on our everyday lives. The world of sport is no exception, with AI rapidly changing the way that sports are played, managed, and analyzed.
Let’s look at a few examples of AI being incorporated into various sports, both on the field and off it.
Game strategy and tactics
The use of real-time data to help with decision-making has been common for some time now, especially in team sports like soccer and basketball. Players wear GPS-enabled devices, usually integrated into their clothing, to show exactly where they’re running on the field of play. They can also measure vital physical statistics such as heart rate and hydration levels, giving up-to-date insights into a player’s condition at any given time.
AI can instantly gather and analyze this data to find patterns that coaches might otherwise miss. Players can be rotated or shifted based on AI-generated suggestions, helping them to find the optimal strategies to break down an opponent’s defense. AI models can also be trained with historical data on opponents’ performance, helping them to predict player movements in-game and accurately forecast their reactions to various tactics.
Sounds a bit…robotic? Don’t worry. AI isn’t anticipated to replace coaches – who bring less quantifiable but still crucial human inputs such as experience and motivation – but supplement them to improve human decision-making and gameplay.
A joint paper by Premier League club Liverpool FC and Google’s AI laboratory Deepmind shares its goal of developing an Automated Video Assistant Coach (AVAC), a system capable of processing raw broadcast video footage and accordingly advising coaching staff in pre, in, and post-match scenarios.
“The AVAC system is an example of what we believe to be the future of human-centric AI research for football, with the aim of integrating all aspects of the frontiers into a cohesive system enabling both understanding and improvement of human football play,” the researchers write.
“Such an AVAC system is envisioned to improve the overall experience of the game for players, coaches, and spectators alike.”
Training and optimizing athletic performance
Off the field of play, AI is being used to train athletes to be the best they can be. Using data gathered from wearable devices and video analysis, AI software can design personalized workout routines that can work with an athlete's strengths, weaknesses, and recovery capabilities. Coaches can use this to provide training sessions that are optimized for maximum results while reducing the risk of injuries and fatigue.
AI’s pattern-reading capabilities are also being used to prevent injuries. Catapult Sports, for example, offers AI-powered athlete monitoring products that allow coaches to track fatigue levels and the potential for injury. Using real-time and personalized historical data, machine learning technology is able to accurately predict the likelihood of an injury and allow coaches to adjust training workloads accordingly.
“Athlete monitoring technology is instrumental in injury risk management,” Catapult Sports write in a report.
“By keeping a watchful eye on athletes’ workloads and biomechanics, the tech helps identify potential overuse or strain, enabling early intervention. This proactive approach to athlete health is a game-changer, significantly reducing downtime and fostering longer, more productive careers.”
Refereeing and fair play
With so much riding on results in elite sport, getting decisions right and ensuring a level playing field is essential. Video-assisted refereeing has been an integral part of sports like tennis, cricket, football, and rugby union for some time and is now being augmented by AI to improve and speed up decision-making.
For example, at the 2022 World Cup, football’s Video-Assisted Referee (VAR) was helped by Semi-Automated Offside Technology (SAOT), an AI system that uses location data to map the positions of players when a ball is kicked. Data is collected from the ball, which broadcasts its location over 500 times per second, and 12 cameras around the stadium that track 29 different points on each player’s body. In the event of a close offside call, all of this data is collated by SAOT to give an accurate offside decision in a matter of seconds. It also generates a 3D image showing the offside line and a player’s position relative to it, which is shown to the on-pitch referee and viewers in the stadium and on TV.
Difficult yet crucial decisions like offside in football, line calls in tennis, and LBW in cricket have long been the source of controversy. They’re almost impossible to get consistently right with the human eye, leaving referees and match officials open to accusations of incompetence and, even worse, corruption.
One sport that’s suffered more than most in this respect is boxing. The judge-based scoring system, which relies on points scored by subjective criteria such as “clean” hits, aggressiveness, and ring generalship, is open to abuse by corrupt judges, and many fights have been marred by obvious favoritism and deep controversy.
Deepstrike – AI for combat sports
To find out more, we spoke to Allan Svejstrup, CEO of Jabbr and creator of Computer Vision AI tools specifically designed for combat sports. Driven by the belief that combat sports need transparency and fairness, Jabbr’s flagship product, Deepstrike, aims to make automatic content generation, stats, analytics, and professional-level streaming available to anyone and everyone.
Svejstrup says that boxing’s struggle with fair and transparent scoring was a primary motivation behind Deepstrike. “I started watching boxing when I was about 13-14 years old, but in the last few years, I've been somewhat frustrated by obvious bout manipulation and the lack of quality analytics in the sport,” he told Cybernews.
His initial idea was to build an AI version of CompuBox, which involves humans watching the fight live and counting punches with clickers in their hands. “I'd been building AI systems for human action-activity recognition for a few years, and I reckoned it would be possible to build an AI that would deliver much more accurate stats with more metrics, as well as cheaper so that it would be available for all fights, not just championship events,” he explained.
Jabbr’s vision is to make automated content creation, streaming, and fight analytics available to everyone in the fight game, no matter how big or small they are. “High-level stats and analytics require dedicated performance analysts to do manual event tagging, which is expensive. We want to make them available to all the fighters and clubs out there at a fraction of the cost,” Svejstrup continued.
The quest to level the playing field and make quality analytics available to all, rather than just an elite few, is admirable, and the reception has been overwhelmingly positive. But what about the question of scoring?
“Lots of the attention coming our way has also been about improving the judging aspect that everyone is so frustrated about,” Svejstrup says. “In theory, future versions of Deepstrike could be used for scoring fights as well, but that's quite political. Different people have different ideas and thoughts on how a fight should be scored, like which aspects are more and less important, for example.”
“I think giving judges access to high-quality live stats could reduce the number of instances of those more absurd decisions, but it probably wouldn't change much when it comes to the more subtle bout manipulation. Commissions and promotions have a strong vested interest in making sure that whichever fighter is expected to bring in more revenue will get preferential treatment by judges.”
So, while AI analytics is already changing sport in more ways than you’d expect, it has its limitations. Human characteristics, both good and bad, still dominate and are the reasons why millions tune in to watch their favorite athletes and teams every day. That said, there’s still plenty of room for improvement, and AI’s role in sport will continue to grow.
“AI-enhanced tools for sports are definitely here to stay. Look at what it's done for other sports, such as soccer,” Svejstrup concludes. “It's just a matter of time before it'll bring about big changes in combat sports as well.”
More from Cybernews
Subscribe to our newsletter