news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Key Points

Tencent Technology News, July 27, according to foreign media reports, as more than 10,000 athletes gathered in Paris to represent about 200 countries and regions to participate in the 2024 Summer Olympics, they encountered a new, friendly but invisible voice companion - AthleteGPT. This artificial intelligence chatbot designed specifically for athletes is integrated into the Athlete365 mobile application to provide athletes with instant guidance and other services.

Athletes only need to ask simple questions such as "How do I get to the competition venue?", "Can it be live broadcast?" or "Will the competition use computer referees?" or even "Where can I get sponsor gifts?" AthleteGPT can quickly integrate massive amounts of information and provide instant responses to ensure that every athlete's questions are answered efficiently.

According to Todd Harple, director of the Olympic AI innovation project at Intel Labs in Hillsboro, Oregon, AthleteGPT's superior information processing capabilities can "quickly retrieve thousands of pages of data and provide athletes with accurate information anytime, anywhere."

This chatbot is essentially still a large language model (LLM), built with French company Mistral AI and Intel’s Gaudi processor, is yet another way AI is leaving its mark at this year’s Olympics.ChatGPTTechnologies such as kinetic energy and speed were still unknown, but the sprinters at the Paris Olympics will undoubtedly experience first-hand the leaps and bounds that have been made in the field since then.

The International Olympic Committee (IOC) is embracing this technological change with an open attitude and released the AI ​​Agenda in April this year, which aims to guide and regulate the booming development of AI in the sports field, while carefully planning its strategic application in the Olympic Games. "We must be the leaders of change, not the objects of change," IOC President Thomas Bach emphasized at a press conference in London, where the cutting-edge functions of several AI sports tools were also demonstrated.

Nature explores three ways AI is reshaping the Olympic experience for athletes and spectators:

1. Gain insights into athlete performance and training

As early as 1900, when Paris hosted the first Olympic Games, French scientist Étienne-Jules Marey pioneered the use of technology to capture sports moments. His high-speed chronograph technology was like a machine gun, freezing the movements of sprinters and long jumpers on film. In 1901, Nature magazine highly praised him, believing that he had analyzed the mysteries of human biomechanics and revealed the competitive advantages of athletes.

Today, technological advances have made smartphones a powerful tool for sports analysis. Intel's 3D Athlete Tracking (3DAT) technology uses artificial intelligence to accurately lock 21 key points on the human body and present the movement trajectory in detail, providing coaches with the same biomechanical insights as elite athletes during training. Harpur predicts that these technologies will not only increase the intensity of the arena, but may also help set new records.

The application of artificial intelligence in improving athlete performance is extensive and in-depth, from customized sports equipment to personalized nutrition and training programs, all of which reflect the power of technology. "This may even accelerate our discovery of new sports strategies," said Harpur. Just like the birth of the revolutionary technique of the Foss high jump, it foreshadows the profound impact and infinite possibilities of technology on sports. This high jump was first created by American athlete Dick Fosbury in the 1968 Olympics and is now called the back jump.

The convenience of personal data collection and the deep integration of artificial intelligence are opening up a new talent identification path for coaches, making the sports competition arena more fair and transparent. In March this year, the International Olympic Committee successfully piloted a cutting-edge scouting program, relying on 3DAT technology, and discovered more than 40 future Olympic stars with unlimited potential in Senegal through detailed analysis of basic training such as running and jumping.

However, it is worth noting that sports powers with large professional leagues and abundant resources enjoy significant advantages in data collection and analysis, and can use massive amounts of high-quality data to continuously optimize training algorithms. In this regard, Patrick Lucey, chief scientist at Stats Perform Sports Technology, pointed out that some Olympic events face the challenge of data scarcity, that is, "limited data footprint." Nevertheless, this situation has also inspired innovative applications of artificial intelligence technology in other key areas such as refereeing.

2. Referees and real-time data

Frank Ohme, an Olympic water polo referee and astrophysicist at the Max Planck Institute for Gravitational Physics in Hannover, Germany, is no stranger to artificial intelligence. In his daily work, he shuttles through complex gravitational wave data, looking for subtle signals of black hole collisions, sometimes with the help of artificial intelligence. However, when he stood by the pool at the Paris Olympics in his white referee uniform, he faced a completely different challenge: accurately judging whether the ball had crossed the goal line in the splashing water.

Although artificial intelligence has supported similar decisions in sports such as football, such as through multiple cameras capturing images of the field and chip technology built into the ball, its popularity in other sports has been slow, especially in the refereeing stage that requires real-time data analysis.

Lack of funds and the individual needs of different events, especially the 32 events in the Paris Olympics, also constitute a major obstacle to the promotion of artificial intelligence technology. Ohm pointed out that although water polo has a long history, as one of the oldest Olympic team events, its economic influence is difficult to compare with popular events such as basketball and football. In addition, the underwater environment and image capture in fierce confrontation also pose unique technical challenges for the application of artificial intelligence in the field of water polo.

In real-time adjudication assisted by artificial intelligence, accurate and transparent communication is essential. Ohm believes that the best way to gain trust and recognition is to fully present information to the team and the audience through images or visualizations, and to empower them to make their own judgments.

As for foul calls in contact sports, the ambiguity is hard to eradicate, and these fast-changing decisions are often difficult to agree on. Ohm added: "Compared to detecting black holes in the universe, quantifying these fouls and handing them over to artificial intelligence is undoubtedly a more complex task."

3. Enhance audience experience

The massive amount of data accumulated during the game can not only provide training support for artificial intelligence algorithms, but also open up new horizons for TV viewers who are hungry for statistical information. Scientist Patrick Lucey is well aware of this. He said: "Sports has its own unique language, which transcends boundaries and promotes global communication. The integration of statistics and data adds rich elements to these exchanges and deepens the level of dialogue, which is exactly what the audience expects."

Faced with this emerging information trend, major broadcasters are intensively exploring how to present it, striving to display more data on the TV screen. At the 2000 Sydney Olympics, when the "virtual record line" was superimposed on the TV screen, the audience was fascinated. Now, looking forward to 2024, Harpur foresees that broadcasters will have unprecedented capabilities to intuitively present more dimensional data such as acceleration, extreme speed, and cadence, bringing unprecedented viewing experience to the audience.

What excites Harpur the most is that with the help of Intel Geti computer vision artificial intelligence platform, the idea of ​​personalized highlights is gradually becoming a reality, which may lead the trend of future broadcasting. Harpur said that in the face of massive event records, artificial intelligence is like a precise sieve that can quickly capture and weave the most desired scenes in the hearts of every audience, completely subverting the traditional viewing mode.

This change will especially benefit coaches and broadcasters in resource-limited countries. "If someone wants to see every three-pointer made by the Nigerian men's basketball team, artificial intelligence can go through all the shots and automatically put them together," said Hapur. (Compiled by Jinlu)