Artificial intelligence is already in football: in the stadium cameras, in refereeing decisions, in scouting departments and in the way fans follow a match. This guide gives you the map of what exists today and what is coming.
What we mean by AI in football
Programs that learn from data to recognise patterns: where each player and the ball are, what kind of action happened, how likely a goal was, or which player profile resembles another. They do not "understand" football like a person: they calculate from what they were taught.
What already exists
- Tracking and video data: cameras and sensors turn a match into position and action data. That is how pressing, the space each team controls or pass maps are measured.
- Semi-automated offside: cameras that track the players and a sensor in the ball help the video referee. It was used at the Qatar 2022 World Cup and in club competitions. The final decision is still human.
- Goal-line technology: camera systems calculate in milliseconds whether the ball crossed the line.
- Expected goals (xG): a model estimates the probability that a shot is a goal, and helps show whether a team played better or worse than the score says.
- Scouting with algorithms: clubs filter thousands of players with data and video, and scouts then watch the ones the system highlights.
- Load and health: GPS vests and watches flag changes in training load. Predicting an injury is much harder than some advertising promises.
- Fan content: summaries, translations and personalised recommendations, and stadiums that manage access and queues with data.
What is coming
These ideas are being tested or are still concepts, and the images on this site that show them are illustrations, not products that exist today:
- Faster, more accessible movement analysis and play prediction, also for small clubs.
- Light sensors in clothing or in the ball that give more data without bothering the player.
- Immersive fan experiences: watching the match with data and personalised replays from the stands or from home.
- Automated stadium services, such as food delivery to the seats.
Limits and responsibility
A model inherits the biases of its data, for example the relative age effect in youth football. Data about minors needs special care. And a prediction is not a promise: AI helps decide, but the coach, the referee, the doctor or the scout are still the ones who decide.
Where to go next
Pick your profile below or open the magazine: short articles on each topic, and you can sign up to get the daily summary.