The Ball, the Avatars, and the New Generation of Semi-Automated Offside Technology
At the heart of the tournament's technological infrastructure lies the evolution of the Semi-Automated Offside Technology (SAOT), a system that combines tracking cameras mounted beneath stadium roofs with real-time data transmitted by the official Adidas Trionda match ball.
At the 2026 FIFA World Cup, every stadium deployed 16 high-resolution cameras dedicated exclusively to offside detection, compared with just 12 at the 2022 tournament in Qatar. These cameras simultaneously tracked up to 29 body points for every player multiple times per second. The Trionda ball—named after the tournament's three host nations—contains an inertial sensor developed in partnership with the German technology company Kinexon, capable of recording the ball's movement 500 times per second and transmitting this data directly to the Video Assistant Referee (VAR) system.
Before the tournament kicked off, each of the 1,248 qualified players underwent a full-body digital scan, from which personalized three-dimensional avatars were created. These digital replicas are capable of reconstructing the precise position of every part of a player's body at the exact moment of a contested play.
By combining data from the sensor-equipped ball with the position of these avatars, the system can now detect offside infringements as small as 10 centimeters, compared with the 50-centimeter threshold used just four years ago. The increase in precision is remarkable—perhaps even greater than what the human eye can naturally perceive or intuitively accept as fair. A difference of only a few centimeters, impossible to detect at the speed of live play, can now be enough to disallow a goal that appeared perfectly legitimate to spectators in the stadium or watching on television.
This tension became evident during a group-stage match when a stoppage-time goal scored by Colombian defender Davinson Sánchez was ruled out for offside. The television broadcast showed viewers the player's own three-dimensional avatar, with the infringement measured at the tip of his boot. Fans disputed the referee's decision, but at that moment no one challenged what the visualization itself revealed. Instead of seeing only a whistle, a raised flag, and a brief explanation, viewers were shown the exact 3D avatar and the virtual offside line on which the decision had been based.
Football AI Pro and the Democratization of Tactical Analysis
Beyond officiating, perhaps the tournament's most ambitious innovation is Football AI Pro, a platform developed jointly by FIFA and Lenovo around a proprietary large language model trained on hundreds of millions of historical data points collected over decades of FIFA competitions.
Unlike previous generations of analytical tools, which were typically available only to organizations with substantial financial resources, Football AI Pro was made available to all 48 participating national teams, regardless of their budget or technological capabilities.
Coaches and analysts could ask questions in natural language, much as they would in an ordinary conversation, receiving answers in the form of written reports, charts, tactical simulations, and animated visual reconstructions of match scenarios—not merely raw statistics.
It represents a genuine democratization of advanced tactical analysis, which until now had largely remained an advantage reserved for football's wealthiest federations, capable of maintaining dedicated data analysis departments.
The tournament also introduced a series of technologies designed to improve both officiating transparency and participant safety. AI-stabilized cameras mounted directly on referees' equipment provided the referee's point of view in all 104 matches of the tournament. At the same time, separate automated moderation systems monitored social media platforms using AI models trained to recognize abusive language patterns across multiple languages, enabling them to identify and block hate speech and harassment directed at players.
How a World Cup Team Trains with the Help of Artificial Intelligence
Long before the first ball was kicked at this year's tournament, artificial intelligence had already begun shaping nearly every stage of national team preparation: from deciding which players are called up, to how they train, and even when they should enter the match.
Many national team coaches now rely on performance analysis platforms fed by data collected from GPS trackers worn during training sessions, combined with biometric sensors that monitor fatigue and recovery, as well as video analysis systems capable of automatically tagging every technical action across thousands of hours of footage. The practical result is that coaching staffs which, until recently, depended largely on manually reviewing recordings and subjective observation now have access to precise, data-driven metrics for every player on a daily basis: how often and how intensely a player presses after losing possession, how effectively they maintain their defensive positioning relative to teammates, or how quickly they recover to a normal heart rate between consecutive sprints.
Predicting injury risk has also become increasingly data-driven. Wearable sensors monitor muscle activation and joint movement, allowing AI systems to recommend individualized training programs tailored to each player's physical condition on any given day. The same models are used during matches to estimate the optimal moment for making substitutions—an advantage that can prove decisive in the closing stages of a game.
At the opposite end of the talent pipeline, long before a player is ever considered for a national team, other AI models are already analyzing video footage from youth tournaments, studying how young footballers move, position themselves on the pitch, and develop physically over time. The goal is to identify patterns of talent that even the most experienced scout might simply overlook because no human can realistically watch thousands of hours of matches from lesser-known competitions.
In an official FIFA statement, President Gianni Infantino described Football AI Pro as a tool designed to democratize access to performance analysis by providing all 48 participating national teams with the same comprehensive analytical dataset. The statement neatly captures the broader ambition behind the platform: reducing the analytical gap between football federations.
The platform does, however, include one important limitation. According to FIFA's official presentation, Football AI Pro cannot be used during live matches. To preserve the integrity of human tactical decision-making, it is restricted to pre-match preparation and post-match analysis.
The coach's role is therefore not disappearing. Rather, it is being redefined—from the person responsible for finding information to the person responsible for deciding what to do with the information that AI has already placed on the table.
Not Everyone Is Applauding
Making the game more transparent inevitably requires a technological infrastructure that is also more capable of tracking people, objects, and behavior within a public space.
The scale of this technological deployment has not gone unnoticed by civil liberties organizations. The implementation of such a dense visual tracking system across 16 stadiums in three different countries has drawn criticism from organizations including the American Civil Liberties Union (ACLU) and Privacy International. Their concern is not directed at football itself, but at the precedent established by deploying large-scale visual recognition and individual tracking technologies in public environments of this magnitude, even if the systems are officially intended to monitor the game rather than the spectators.
Club Football Had Already Been Testing These Technologies
Most of the technologies showcased at this year's World Cup did not emerge overnight. They have already been tested and refined for years in elite club competitions such as the Premier League, the UEFA Champions League, and La Liga.
Professional clubs routinely use specialized platforms such as those developed by Stats Perform and SciSports to automate scouting in smaller, less-publicized leagues where traditional scouts rarely have the opportunity to travel.
Machine learning algorithms—a branch of artificial intelligence capable of discovering patterns directly from large datasets without being explicitly programmed for every individual scenario—simulate thousands of possible versions of the same match before kickoff, helping coaching staffs determine optimal formations and set-piece strategies.
At the same time, GPS-enabled performance vests worn during training sessions continuously monitor heart rate, distance covered, and sprint intensity, alerting medical staff before accumulated fatigue develops into a muscular injury.
Precision Gained, Emotion Lost
Automation has been introduced precisely where both the question and the answer can be measured objectively and reduced to clear, verifiable data: the position of a player's body at a specific fraction of a second, the speed of a sprint, or the trajectory of a ball. Human judgment remains essential wherever interpretation is required—whether an offside player actually influenced the play or whether a reckless tackle deserves a yellow or a red card. In practice, this has become a carefully balanced division of labor between sensors and human officials.
There is, however, a price to this increased precision that statistics alone cannot capture: the way AI-assisted officiating changes the emotional experience of being a football fan.
A goal is almost never an immediately completed moment anymore. It is increasingly followed by pauses lasting tens of seconds—or sometimes several minutes—during which the eruption of joy inside the stadium remains suspended while the virtual review determines whether the goal will ultimately stand.
The replay itself, presented through a three-dimensional player avatar, undoubtedly makes the referee's decision easier to understand. Yet it also transforms what was once a spontaneous emotional moment—the celebration, the embrace, the collective explosion of excitement—into a technical validation process examined frame by frame before anyone is truly allowed to celebrate.
This fragmentation of the game, with repeated interruptions that can sometimes last longer than the attacking move itself, represents the inevitable trade-off for greater accuracy. Football becomes fairer, but part of the continuous, almost uncontrollable emotional intensity that has always defined the sport is inevitably sacrificed.
It is therefore no coincidence that many of the criticisms voiced by supporters throughout the tournament were not directed at the accuracy of AI-assisted decisions. Instead, they focused on the dead time introduced into the game—the dramatic pause that turns spontaneous joy into a celebration conditional upon a technical verdict.
The 2026 FIFA World Cup appears to have demonstrated that football can absorb a remarkable amount of artificial intelligence without losing its identity as a sport ultimately officiated by humans. Yet the precise point at which this balance ceases to be acceptable for fans, players, and officials alike has not yet been fully tested. As AI continues to advance, technological momentum will inevitably push more and more decisions that still belong to human judgment toward automation.
