How Is Artificial Intelligence Changing Football?

0:00 if you want to be the best in professional football you're going to have to need the best players the best training and the best strategy and now to stay at the top of the game it's becoming increasingly important to also have the best AI so how is artificial intelligence changing football could you imagine supporting a football team made up of robots these robots are already able to play against each other on the football field the goal of the robo cup organizers is that by 20 50 their robots will be able to defeat football's world champions to be honest though it's hard to imagine traditional football fans being that excited about a robot football team but technology is becoming more and more important in sport whether fans like it

0:45 or not a digital clock for documenting yellow and red cards Lions people keep the ref in the loop via radio and with the help of video assistant referees or V the game is assessed from all angles pitchside monitors provide video evidence it's undeniable professional football is more technical than ever yandre is responsible for referee Innovation at the German Football Association known as the dayf bay the use of Technology makes soccer fairer it makes decisions verifiable more transparent decisions can be assessed as right or wrong the vast majority of decisions can be categorized in this

1:32 way but fans aren't always as enthusiastic it can take a long time for dat to be evaluated you have to be careful that the technology doesn't take over you have to differentiate between times when using it makes sense and when it might endanger the nature of the game at the end of the day it's still the referee who must decide whether the outputs provided match their perception on the pitch and that's where AI applications come in the world Football Association FIFA has already gradually used a new semi-automatic offside detector in tournaments cameras mounted under the stadium roof and sensors in the ball send data to an AI system this then

2:20 calculates the relative position of players in the ball 500 times a second it then generates 3D models and notifies the referee when a player is offside [Music] someone calls it robot offside it's not still the referees are the assistant referees still responsible for the decision on the field of play V is a controversial Topic in football but it's not going anywhere so football fans might as well embrace it and that's where artificial intelligence can actually help as the tech gets better referees will be able to make better decisions of course that means training the systems properly and football authorities do need to ensure it's safeguarded against manipulation by

3:07 humans there have been referees who have deliberately manipulated matches so of course when introducing Technologies we must ensure that they run as securely as possible for example that they cannot be hacked or manipulated from the outside in order for AI models to work they need a lot of data luckily football already has bucket loads in modern game analysis there are two data types there's event data like who passes how often and then there's positional data Professor memt from the German sport University cologne explains here we see positional data in both football and handball the video and the corresponding visualized position

3:53 data individual players Behavior can be extrapolated from positional data tactical structure of the team can also be recognized and heat Maps show who was where during the game the data is gathered by trackers in chest straps worn by the players receivers around the pitch ensure that the data is transmitted securely the stats are then analyzed in real time online another way of collecting positional data is to evaluate images using so-called computer vision but one difficulty is that not all players are always visible in the image AI could be the solution here too the more data you have the higher

4:40 the probability that you will find patterns those patterns can be used to establish a correlation which can help predict future outcomes at the end of the day football is complex many parameters play a role but the more parameters I include in my analysis the closer I am to reality the importance of data is increasing as is the demand for match analysts they're often considered just as important as traditional assistant coaches but to be able to use the data sensibly analysts need training most match analysts don't have the programming skills needed for working with data in a targeted and quick fashion so far comprehensive data

5:26 collection exists mostly in top European leagues but in the future A team's success will depend not only on having the best players and coaches but also the best programmers now it's all about who builds the best AI who has the best parameters and who can use outputs from games or training they must draw the right conclusions communicate Essentials to the players and ultimately improve training programs analyzing data as quickly as possible with AI has become crucial in professional football and it's even able to predict the probability of injuries England International Harry Kane has invested in an AI system which uses

6:12 real-time data to maximize training programs this can prevent injuries and optimize performance artificial intelligence can also detect exhaustion which could help coaches during matches they'd be able to take players off the field before they even have a chance of getting injured this is just just one area where coaches could make use of AI and the tech is already helping them to find weaknesses in the opposition typically match analysis takes place on screen like here at the German sport University cologne data is collected and analyzed such as this opponent's behavior of corner kicks for example precise but timec consuming analyzing a game with video data would take us about 8 hours

6:59 depending on how many parameters we include if we have positioned data and analyze it with our software the same task takes 8 seconds so we can evaluate a lot more games in a shorter time a clear Advantage for those teams already working with AI those who can evaluate more games will more readily recognize patterns in opponent's tactics and based on that training plans can be tailored for each new opponent Daniel memt says the effectiveness of a new Tactical approach can be tested too he and his team have access to data from the German [Applause] [Music]

7:47 Bundesliga you can test new group tactics or new skills beforehand we could test those things here at the sports University and say the next day whether it made a difference or not until today such insights would become apparent only over the course of a season or even after a few defeats but now pre-season AI analysis allows for more targeted training tactics that the AI deems unsuccessful are avoided but for Daniel mem football success can't only be planned on a computer football differs from many other sports in that there is a large element of chance we're talking 42% of

8:34 all goals scored in unplanned actions that can be trained and that's why the game will always remain a bit mystical maybe that's why it's so exciting football will always have that unpredictable element which makes it so loved by millions across the world but algorithms could still help with many important questions like how can players improve their game even further or is it is it worth buying a certain player that's what German analysis platform player is attempting to achieve with their [Music] system data gives Talent visibility and the more data we have the more visible Global Talent will be for us personally

9:20 I believe that AI will completely change football and democratized football able data becomes the easier it will be for smaller clubs who don't have 20 50 head scouting teams to find players for their ranks the AI platform player analyzes data to find the next generation of football stars making recommendations to clubs on who to look out for player uses data from football associations but also collects its own we also note whether someone scores a 1 nil or a 4-nil whether someone passes

10:06 forward or backwards whether it happens Under Pressure whether it triggers something whether it influences the rest of the game this involves 400 data points which we have converted into 12 Mega matrices 12 data points that are most important for a particular position today player compiles and Compares data from close to a mill million football games what's more 300,000 players are documented in their system such as Bayern Munich's Jamal M I can see what his potential is and as a club I can plan for that I see 2024 is

10:51 his year after that he'll lose a little momentum but not significantly in 2025 2026 he'll continue to play at the top of his [Music] game the makers claim the AI can calculate players future performance with an impressive 90% accuracy to further improve the AI forecasts are frequently monitored and results scrutinized could this mean goodbye for its human counterparts um the question implies a kind of fear a fear that machines will take over football and everything else too I don't

11:36 believe that will happen it'll never happen with progress it always depends on whether we can approach Innovation fearlessly bravely and wisely to make better [Music] decisions embracing Tech can only be beneficial for football unfortunately these systems still aren't available for everyone some of them such as player are currently only focusing on the men's game and it's clear that to use AI in football at the moment you need cash so will artificial intelligence actually make the game better or will it just make the rich Club stronger let us know what you think and see you next time

12:23 [Music]