Football 3.0: The beautiful game in the age of Artificial Intelligence
Football 1.0 remained pure and humane, Football 2.0 could be considered what the sport looks like today, with major technological innovations fitting into it. But how will the beautiful game look like as our AI systems become more intelligent and applicable to the world of football?
At the time of Stable Diffusion and ChatGPT, the beautiful game has not remained apathetic to the current significant technological advancements. Technology and Artificial Intelligence have not kept their hand away from the king of sports. New innovations attempt to contribute to the fairness and beauty of the game, albeit the greatest facelifting of football might still be on hold.
It’s already been more than a month since we witnessed the greatest world cup final, and one of the greatest stories in football history being artistically and breathtakingly completed. And as controversial as the whole Qatar scenery might have been, people’s love for the game could not be restrained. While the celebrations in Rosario, Santa Fe wane, it’s worth shooting a glance at the technology that was utilized at the tournament, and in general, how our scientific breakthroughs find their way into the world of football.
World Cup 2022
Semi-automated offside technology
Except for the retractable roof, the Al Bayt stadium at Al Khor, Qatar, on 20 November 2022 when it hosted the first game of the tournament, also maintained twelve cameras mounted underneath its roof. Those dedicated tracking cameras were used to provide real-time data to support the decisions of the officials. Utilizing state-of-the-art computer vision methods, the cameras could track the ball and 29 data points of the body of each player 50 times per second, a technique widely known as Human Pose Estimation. Additionally, an inertial measurement unit (IMU) sensor was placed inside the Al Rihla (the name of the ball — means ‘the journey’ in Arabic). This sensor, positioned in the center of the ball, could send ball data to the video operation room 500 times per second, leaving no margin for error in detecting the precise kick point. The entire system was capable of communicating automated offside alerts, and in combination with 3D animations allowed for improved and fairly quick decisions.
Stadiums’ Digital Twins
The stadiums in Qatar were also equipped with myriads of sensors that were used to facilitate the control of the temperature (and not only). The venues also featured solar-power cooling systems, and in a combination with approximately 40,000 IoT devices, accommodated centralized management of multiple essential operations such as climate control and security. Impressively, the whole infrastructure is evidence of our engineering and scientific capabilities and the ingenuity to work in the intersection of disciplines. The digital twin technology is a creation emergent from the collaboration of multiple scientific areas: Internet of Things (IoT), Software Engineering, Cloud Computing, Computer Vision, and certainly I miss a few.
FIFA Player App
Similarly to our smartwatches providing us with daily activity reports, tracking our steps and our evening runs, a Player App developed by FIFA granted the players the luxury of individual football data metrics. The explosion of data and its availability made it possible for physical performance metrics and enhanced football intelligence metrics generated through an array of developed algorithms and artificial intelligence models.
Besides Qatar
Of course, world-class AI research could not dismiss the vast amounts of data and the potential of sports analytics. Deepmind acknowledged that a reciprocal dialogue between AI and football should be established, as it can be significantly beneficial for both football and AI research. Combining video understanding, statistical learning, and game theory, they envision a radically transformed game — and other sports could also follow.
Technology and AI sought and found applicability to a number of different processes and operations in the world of football. Smart portable AI cameras are capable of recording, streaming, and analyzing football matches (Veo), while Scouters can make informed decisions about players’ analyzed performances with the use of AI-powered Scouting platforms (AIScout).
Football 3.0
The title of this article is inspired by Max Tegmark’s work: Life 3.0. The famous cosmologist discusses various implications and potential futures for humanity, as well as multiple scenarios that our societies could face in the age of emerging technologies and artificial intelligence.
Similarly, while Football 1.0 remained true to its origins and purity, Football 2.0 could be considered what the sport looks like today: Goal-line technologies, VARs, sports analytics, intelligent scouting, etc. The beautiful game has already faced a transformation. And while the scientific world promises and prepares for Artificial General Intelligence — or human-level intelligence —, the world of football must be ready to welcome new technologies without taking away the humanness and emotion that define the game.
“Ladies and gentlemen, this is the moment we’ve all been waiting for. Lionel Messi, captain of Argentina, lifts the World Cup in triumph. The emotion on his face, the roar of the crowd, all a testament to this historic moment. Years of hard work, sacrifices, and unwavering determination have led to this. Messi, a true legend of the sport, has once again proven why he is the best. The World Cup is his and the world is in awe. This is a moment that will live forever in the annals of football history.”
Yes, I’ve asked ChatGPT to generate a commentary in the style of Peter Drury. If you are familiar with his iconic and poetic style, you could admit that this is not too far from it. And that’s already here. Text-to-speech technology is also here. What if in the near future, we witness football events being live commentated by an AI technology?
We’ve already created systems that can replicate or reproduce the style of Van Gogh in generated images and videos. What if we create systems capable of capturing and reproducing the style of individual players? What if our simulations become not only realistic and indistinguishable from the real world, but also learn to produce and track the fans’ emotions? What if large models trained on vast amounts of data, learn to identify the key moments that amplify the spectators’ emotions and reproduce that on photorealistic humanoid football simulations? Okay, this may sound too extreme and possibly undesired.
But maybe we are not far off using technology to provide us with the ground truth decisions regarding a football match’s refereeing. Maybe the whistles could be handed to systems powered by AI, which are much more reliable than human eyes. Advancing the semi-automated offside technology, enhancing it with more cameras, and training computer vision models to make real-time decisions based on the rules of the game, might reduce some of the controversies, increase the fairness of the game, and even eliminate any wasted time on difficult events. Our algorithms and models along with the availability of data can potentially lead to world-class and even near-perfect (if not perfect) refereeing. And if replacing human referees with intelligent systems means that human error is substantially reduced while decisions are not delayed, we would be naive not to admit that we will be okay if we don’t hear the boos of fans towards poor (or maybe not) referees’ decisions.
Ultimately, it’s about the beautiful game. Unquestionably, employing artificial intelligence for its flourishing encloses some dangers. And we should identify and address those while we attempt to transform the sport. Most importantly, we should keep it intact from automations that can potentially dehumanize or negatively impact the fans' or players’ experiences. Arguably, though, new technologies and AI are capable of elevating the sport and contributing to its beauty.
