Next-generation semi-automated offside technology (SAOT) is applied for the first time at this World Cup /Courtesy of FIFA website

The World Cup, the world's biggest soccer festival, kicked off on the 11th of this month. The final of the 2026 Fédération Internationale de Football Association (FIFA) North and Central America World Cup, co-hosted by the United States, Canada and Mexico, is expected to draw about 1.5 billion viewers. With countless fans watching players' passes, touches and goals, FIFA is applying advanced artificial intelligence (AI) technology to matches for faster and more accurate decisions.

FIFA introduced next-generation "semi-automated offside technology (SAOT)" for the first time at this World Cup. Based on players' limb movements and the ball's position during play, AI determines offside in real time and automatically sends an alert to the assistant referee in clear-cut cases.

In the 2022 Qatar World Cup, the existing SAOT first sent AI analysis results to the video assistant referee (VAR), which were then relayed to the assistant referee on the field. In this tournament, when a clear offside is detected, AI sends a voice alert directly to the assistant referee's in-ear: "Offside, offside, offside."

The final decision rests with the assistant referee, not AI, but the assistant can refer to the on-field situation and the real-time AI alert to decide immediately whether to raise the flag, which FIFA says can cut decision wait times and reduce unnecessary play and scuffles between players in the meantime. Johannes Holzmueller, FIFA's director of football technology and innovation, said, "With the assistant referee able to confirm offside in real time and raise the flag, decision speed has become much faster."

To boost the accuracy of calls by the latest SAOT, FIFA has deployed a range of AI-based match analysis tools, including precision cameras and sensors and Deep Learning-based Computer Vision (video analysis) technology. First, 16 cameras installed in each stadium track the movements of the ball and players. Dedicated cameras using the optical tracking technology of Sony subsidiary Hawk-Eye Innovations collect more than 150 million data points per match, 50 times per second, to reconstruct match scenes in 3D (dimensions). The generated 3D views are sent to the VAR officials.

The official match ball for the 2026 FIFA World Cup, Adidas Trinidade /Courtesy of FIFA website

The official World Cup match ball, "Trionda (TRIONDA)," incorporates Adidas' "Connected Ball" technology. An ultra-small inertial measurement unit (IMU) sensor chip embedded inside the ball measures fine movements 500 times per second, including the moment a player strikes the ball and the ball's speed, spin and trajectory. FIFA said, "We transmit this precise data to the VAR system in real time to support the officiating crew's decision-making," adding, "It is particularly helpful in improving the accuracy of offside calls." The ball can capture subtle contacts that are hard to distinguish with the naked eye, and is also used for handball fouls and penalty kick decisions.

FIFA also applied AI-based 3D player avatar technology to SAOT for the first time in partnership with Lenovo. Before the tournament began, 1,248 players from the 48 participating countries underwent precise scans to measure body dimensions. They created "digital avatars" that replicate each player's appearance, including height, arm and leg length, shoulder width and foot size. This allows officials to see at a glance which player was involved in an offside in situations where multiple players overlap or move quickly.

In addition, FIFA is using various review technologies, including Hawk-Eye's VAR, goal-line technology (GLT), and "last touch," which identifies the player who last touched the ball in corner-kick and goal-kick situations.

Chenliang Xu, a professor at the University of Rochester, said, "The technology applied at this World Cup would have been hard to imagine just 10 years ago," explaining that advances in Deep Learning neural networks and graphics processing units (GPUs) have dramatically improved performance in image recognition, object tracking and unstructured data analysis, making it possible to implement advanced sports AI technology.

※ This article has been translated by AI. Share your feedback here.