Player, Manager or Fan – AI Is Delivering Great Improvements in Sport
Sport raises pulses, with spectators rooting for their team in everything from soccer and basketball, to golf and motor racing. Technology has always played a significant role in providing an edge. New materials, improved scientific analysis of motion, and advancements in training techniques continue to push the boundaries in every sporting arena. What will AI be able to add, if anything, to sporting prowess and how we experience and enjoy it?
Perhaps surprising, but the area where AI is having an immediate impact in is the interaction with the spectators and fans. A chatbot using AI has been integrated into the Facebook Messenger interface of sports clubs such as the Sacramento Kings, a Californian basketball team. Fans can pose questions on the team’s lineup, statistics, and even the team’s home arena, with the chatbot responding appropriately. Obviously, a team is only as good as its players. But a supportive, engaged base of fans goes a long way to delivering the edge at games where the odds are more even.
Tearing along at speeds in excess of 200mph, NASCAR delivers thrills and its fair share of spills. Despite improvements in technology and safety, the sport still suffers on average one death annually. Here the Ford Motor Company has invested in AI to help quickly locate race cars on the track. The challenge is heightened by the speed the vehicles are moving, making the available images challenging to classify. The AI was shown to be faster at identifying the desired target vehicle than human members of the team. As a result, the video feed can be used to quickly review any possible damages to the vehicle that might lead to a more serious problem on the track.
Team managers also need help with the big game. Studies have shown that humans place too much weight on a streak of successful gameplay. As a result, players tend to pass to them more often, thinking that they are somehow ‘in the zone’. The reality, as shown in the study “The hot hand in basketball”, is that players overplay another player’s winning streak. Using AI deep learning techniques to analyze games, it is possible to review how players move during games and shed light on the best matchups between players. Managers can use this to take a scientific approach to coaching players, rather than relying on the less-reliable gut feeling.
Even the snippets of the game we see in the highlights round-up is being defined by AI. The team at Wimbledon, the world’s oldest tennis tournament, is using the technique to automatically choose the best bits of the game for television coverage. Drawing upon crowd noise, match data, and the facial expressions of players, the system curates match highlights, saving hours of time in the video editing room.
Of course, AI requires data and lots of it. With changes in the law around handling personal data, especially with GDPR in Europe, there are often challenges in sharing data that can be linked back to players. But these challenges are not new, with coaches, dieticians, scientists and doctors reluctant to share their data amongst one another, even when working for the same team.
On the track, off the pitch, online and during the game – wherever you look, AI is proving itself to be an exceptionally capable tool across all aspects of sport.