The world of sport has always looked to technology to enhance the matchday experience. Now artificial intelligence (AI) is taking it to a whole new level.
Whether you’re an athlete looking for data to give your performance an edge, a coach studying algorithms to fine-tune a game plan, or a referee seeking to always make the right decisions, AI now plays a key role in stadiums and locker rooms. It’s also changing the fan experience, both for those watching live in the stadium or on TV screens at home.
In this article, we examine how sport is growing and changing in the age of AI, and look at what the future holds as technology becomes even more sophisticated – whilst wrestling with the reality that it’s the human factor that creates the headlines and makes sport so dramatic.
So what do sports organisations, athletes, referees and even TV broadcasters need to take into account when considering implementing AI?
Why use AI in Sport?
Sports are a major source of entertainment for both participants and spectators. For example, it’s estimated that more than half of the world’s population aged above four years old follows at least some of the official broadcast coverage of the FIFA World Cup.
The widespread popularity of sports all over the world means it has become a highly profitable industry, attracting huge investment and massive advertising campaigns. And sport has quickly incorporated AI technologies for very clear reasons.
For starters, our shared love of sports creates a lot of business opportunities, and AI technologies can be used to take full advantage of them. But AI in sports goes much deeper than this because of the quality and variety of data available.
Data, Data, and More Data
Think for a moment about the dynamic nature of sports – the way athletes are constantly moving, jumping, running, turning, throwing and kicking. Consider also the trajectory that balls and pucks take, or the speed of a racket head or golf club as it strikes the ball.
Then there’s cardiovascular data, too. We can measure how much energy the athletes have used in their activity, their heart rate, and how much distance they have covered. All of these activities offer a rich source of data that can be recorded through motion capturing and other sensors.
Sporting events also produce a vast amount of heterogeneous data, ranging from match results and player statistics to high-definition videos and visual tracking information which can all be used to train AI and machine learning algorithms. The abundance and variety of this data has led to the development of many successful applications in this field, and the popularity of sports means it is a common area where we can all see AI in action.
Let’s now take a closer look at the existing and potential applications of AI in sports, both on and off the field.
How AI is changing the sports fan experience!
Put simply, AI is getting sports fans – whether they are at the stadium or watching on television – closer to the action and more connected to the game.
Broadcasting companies and producers are now using AI systems to deliver automated, personalized, and more engaging storytelling. Instead of manually annotating each individual play, current AI is able to automatically compile the most interesting highlights in real time by analyzing player movements and crowd atmosphere. This means engagement on social media can happen almost instantly and interactions on social platforms can then be scrutinized in order to improve and personalize future content.
What’s more, AI is now used to provide in-depth analysis and commentary. For example, the shot accuracy of each player at any second during a basketball game can be visualized in a matter of seconds. AI systems give viewers more choice, such as controlling how they follow a particular game by giving them the chance to track their favorite players or even choosing different camera angles.
Media outlets are also using AI to enhance the fan experience and are using speech-to-text 1 technologies on commentators’ voices in order to provide subtitles in different languages. Several AI start-ups are now working on creating holograms of players in stadiums to give spectators a more engaging experience using mixed reality.
Enhancing the Stadium Experience
But it’s not just the fans at home who are benefiting. AI technology is now used in stadiums to generate real-time analytics and virtual replays, provide smart ticketing services, improve crowd management, and create more efficient parking systems. Additionally, AI is now deployed to boost in-stadium security and provide early warning systems through rapid body scan and facial recognition.
Sports events may soon mirror sci-fi movies – many companies are designing drones and robots to deliver snacks to hungry spectators, plus help with cleaning and maintenance. Advertising companies are also working on sentiment analysis of their audiences in order to provide more personalized ads.
How AI is helping players to improve their performance and health.
Fitness and well-being are crucial elements of any sport, and AI-enabled wearable technologies2 are loved by both professional and amateur players looking to monitor their performance and maintain their health.
Smart bracelets, inertial sensors, local and global positioning systems, connected sneakers, smart clothing, and heart rate monitors are some of the most popular. All of these have one thing in common – they give athletes vital data by tracking movement, and they record various physical aspects during workouts, training sessions, and games.
These devices can cover many aspects of performance, from the distance a runner or footballer has covered, to calories burned and beyond. They can also highlight microscopic variations in more physical sports such as boxing. But these devices can do more than just provide statistics – they can also offer athletes recommendations for their training schedules and help them customize their diets with bespoke nutritional plans.
Such analysis helps athletes maintain top-level physical and mental performance, peak training efficiency, and even improve performance during games. It also helps them to detect early signs of fatigue and stress, and prevent musculoskeletal injuries and cardiovascular issues.
Meanwhile, visual tracking systems allow players to identify their key strengths and weaknesses. For example, a footballer can monitor their movements, shots, dribbles and passes, and then work on their skill set to improve their performance in future games.
How AI is enabling teams to recruit the best talent and win more.
The 2011 Hollywood movie Moneyball, starring Brad Pitt and Jonah Hill, tells the story of how a baseball team in the US was one of the first sports outfits to use analytics.
Back in 2002, the Oakland Athletics harnessed data to assemble a competitive team on a small budget. They used what in baseball is called a sabermetric approach3 to measure players’ stats during games. The Athletics were able to compete with some of Major League Baseball’s top teams, such as the New York Yankees, whose wage bill was more than double that of Oakland’s.
At the time, the data used by Oakland was regarded as outlandish and a break from tradition. But while it was groundbreaking, machine intelligence was limited back then, and most decisions were made by humans. However, recent developments in AI and machine learning have brought a whole new level of analysis to the sport.
Now AI can uncover new insights and define successful game day strategies, as well as help teams hire the most qualified staff. In the future, such systems are expected to have a deep understanding of many sophisticated sports – even better than what a human expert has.
The Complexity of Data in Sports
Sport creates a vast amount of data – so much so that it can be hard for even the most trained eye to process all possible insights.
Even when it comes to simple stats from video content, such as the number of passes made by a player in a football match or the distance a rugby player covers during a game, the human eye can’t compete with AI. What’s more, games often feature complex movements and sequences involving several players, something that can’t be quantified using traditional, rule-based algorithms.
One example of this is basketball’s pick-and-roll move. It’s an offensive play involving four players and, naturally, the end result varies. But all pick-and-rolls have one thing in common: they follow a pattern, and identifying patterns is one of AI’s many strengths. Every single metric can now be examined, from player positions on the court, to the speed at which the ball travels.
Teams and coaches have been using tracking technology in the NBA for several years now, and this close examination of all aspects of a player’s game means they can identify young talent, and predict the potential of new players before investing in them. Moreover, by harvesting this data, AI is now able to help coaches improve their game strategies. They can analyze common mistakes, identify strengths and weaknesses in their own team, and target trends and inconsistencies in their opponents.
Many of football’s top clubs are now using AI to make them more competitive in the transfer market. The traditional method of sending scouts to games to watch players is now being complemented by new tech that inserts those reports into dashboards. That data then helps coaches make more educated decisions about their next big-name signing.
AI is also being increasingly used for tactics. Sports teams are no longer putting square pegs into round holes, as modern technology gives insights into a player’s best position on the pitch. Even some of sport’s most experienced coaches, such as Ron Rivera, head coach of the NFL’s Washington Football Team, are using machine learning algorithms to build better strategies to win.
In some cases, AI is noticing things coaches have missed. In basketball, for example, an AI system can now calculate the chances a specific player has of scoring, and will also take into consideration angles and the position of any surrounding defenders. It’s thanks to these advancements why many experts believe a complete AI system, which could fully comprehend a game, might be just around the corner.
How AI is helping referees to make the right decisions.
Referees and umpires are under just as much pressure as coaches and players, and human error is always a possibility. Who can forget Diego Maradona’s infamous “hand of God” handball goal at the 1986 FIFA World Cup Finals?
Recent years, though, have seen the introduction of several systems to help match officials, such as the Decision Review System (DRS) in cricket and the Video Assistant Referee (VAR) in football. But these are based on classic slow-motion and Hawk-Eye4 technology, and still require a human eye to make the final decision. Not only does this break up the flow of a game, it has caused considerable debate and disagreements, with fans, players and commentators all often claiming that VAR gets it wrong.
But several new technologies have been introduced to deliver faster and more accurate decisions that use high resolution sensors to detect the exact placement and speed of a shot or a player. Tennis, for example, now uses an electronic umpire called In/Out, which shows its human counterpart whether a ball has landed out of bounds. Goal-line technology has also become an important part of football matches and a key part of the VAR experience.
Finally, AI is also joining the fight against crime in sport, such as match fixing and betting-related fraud. This technology can analyze sports betting all over the world and identify anomalies and patterns that are unusual, but could still be missed by the human eye.
The Way Forward for AI in Sport
The human factor is what brings excitement and surprises to the world of sport – and that is something AI can never compete with. Players and referees aren’t perfect, but it is those imperfections that are often the main talking points for fans on matchday. Indeed, sport would quickly become pretty boring if it was played in an analytical and predictable fashion.
So it is important to consider that not all elements of sport can be easily quantified, meaning a balance will always be needed between man and machine.
Technologies that provide automatic transcription of audio into written text. ↩
Clothing that incorporates sensors and computer chips to collect and directly process data from the user and their environment. ↩
Sabermetrics is a statistical approach to baseball. The word finds its origin and construction in the acronym SABR (Society for American Baseball Research) which is the first society with the main objective of studying baseball. Read more about it. ↩