At a trade fair stand of the technology group Fujitsu, an amateur basketball player works on her jump shot with the help of artificial intelligence.
Photo: dpa/Shun Kato
In a noisy exhibition hall in Tokyo, Yosuke Nagai throws a basketball past the hoop. A moment later he stands there and listens to an automatic voice telling him how he could have done it better. The 43-year-old amateur athlete should follow Yuki Miyazawa’s example, it is said, while a video wall shows the basketball player from the Japanese first division team Fujitsu Red Wave – and shows how Miyazawa effortlessly throws the ball into the basket with smooth movements.
What did the professional player Miyazawa do differently than the amateur Yosuke Nagai? The computer voice explains: Unlike Nagai, Miyazawa had an angle of 15 degrees between his body axis and his throwing arm when releasing the ball. And she didn’t just throw with her arm, but with full body tension. Yosuke Nagai nods happily when he hears this. »It helps when you see your own posture and then correct it straight away. Of course, when I used to train as a child, there was nothing like that.
Fujitsu promises immediate feedback
His school teacher also explained that you should keep your body tense. “But I couldn’t imagine much of it back then.” Now that could change, Nagai believes. And many of the visitors to the CEATEC trade fair seem to see it similarly. At Japan’s largest technology trade fair, this stand, where you can have your basketball skills tested, is particularly popular. Here the Fujitsu company shows one of its latest developments, which it wants to monetize in the future.
Fujitsu is targeting the areas of health and sports. There is particularly great potential for artificial intelligence (AI) here, explains company spokesman Hidetoshi Tomisaka on the sidelines of the exhibition. »In competitive sports, there have been video analyzes for training so far. But our AI works very quickly. You don’t get your feedback after hours of manual evaluation like before, but almost immediately.«
nd.DieWoche – our weekly newsletter
With our weekly newsletter nd.DailyWords look at the most important topics of the week and read them Highlights our Saturday edition on Friday. Get your free subscription here.
In addition, a normal smartphone camera is sufficient to use the software. “So expensive equipment is no longer necessary.” Tomisaka gives an indication of who Fujitsu is aiming for with its development: it’s about school sports, young talent in competitive sports, but also the professional sector. Because movement sequences can be optimized everywhere. In the first women’s basketball league in Japan, Fujitsu is already using the system for the company’s own team Red Wave.
AI works particularly well in gymnastics
Artificial intelligence has been spreading in many areas of society for several years. She is also already represented in sports. In football, for example, English club Brighton & Hove Albion is known for using AI for its scouting and match analysis. The more data, the better the AI. With its Judging Support System, Fujitsu has also been supporting the work of jurors at World Gymnastics Federation tournaments since 2019 – including the World Championships.
Hidenori Fujiwara, who heads the “Human Digital Twin” department at Fujitsu, sees this as a breakthrough for the use of AI in sports: “In gymnastics, our AI can actually evaluate everything: it recognizes the name of a figure performed, its level of difficulty and the quality the execution.” By having the AI evaluate objectively and not make any mistakes, the audience can also better understand why one person gets more points than the other.
The Japanese technology group senses a billion-dollar market. But how promising is the use of AI in sports really? Daniel Memmert, Professor of Training Science at the German Sport University Cologne, sees a lot of potential, especially in sports in which a jury evaluates the performance. Because people make mistakes: “There are actually nice studies in the field of sports psychology that show that we have a calibration effect.”
At the beginning of a series of evaluations, the evaluation is usually done in the middle. In other words: Since a jury can hardly estimate well at the beginning how good the best performance in the field of participants will be, it tends not to award top marks right from the start. However, an AI would not need this comparison with the other participants for its assessments.
Ideal technology vs. human diversity
When it comes to training, Memmert sees fewer opportunities for AI: “If we look at the technical processes, then you first have to know that there is no ideal technology.” Across all sports, the best athletes in the world are not able to control their own movements to replicate 100 percent in the same situation. »People are good at adapting to new situations, but not at repeating the same situations.«
Memmert is skeptical about AI-based training that supposedly suggests ideal movements: “I don’t even know if that’s a good idea: because what is the gold standard?” Everyone has different body measurements, is different heights, has different abilities different muscle compositions. “People are very, very diverse.”
So will the use of AI be limited to sports that use judges? At Fujitsu you don’t believe that. Hidetoshi Fujiwara says: »We are already looking for ideal types. But we also want to be flexible. With the increase in data availability, this can be achieved.” Discussions are currently being held with clubs in the US baseball league MLB, among other things, to reduce the risk of injury when hit. »We are also active in football.«
And if the principle of AI does not become established in professional sports, what remains is youth and amateur sports, where those practicing often still have to learn very basic things.
sbobet sbobet88 sbobet88 sbobet88