The Financial Times reveals that Liverpool FC is teaming up with Google DeepMind to experiment with artificial intelligence (AI) coaching technology in football.
This season, Liverpool has been scoring a lot of goals from set-piece situations.
DeepMind, the AI research and development division of Google, has previously utilized algorithms to defeat challenging games, stating that patterns in sports could be beneficial for applying AI in automation and traffic regulation.
DeepMind’s AI coaching system, named TacticAI, trained on data from 7,176 corner kicks in the Premier League from 2020 to 2023, is used to provide suggestions to improve player positioning in corner kick situations.
According to a study published in Nature Communications, the majority of suggestions from the computer system were agreed upon by human experts.
Liverpool has refused to confirm whether they have used AI suggestions to improve the effectiveness of set-piece situations or not.
“From an AI perspective, football is a highly dynamic game with many unobservable factors that can influence the outcome. It’s truly a challenging issue,” said Petar Veličković, a DeepMind researcher and co-author of the article in Nature.
On the pitch, the TacticAI system demonstrates both the current capabilities and limitations of efforts to use AI to achieve high performance in sports, surpassing all existing data analysis methods.
This technology holds promise in planning for predictable starting point situations, known as set-pieces, such as corner kicks.
According to the FT, the project DeepMind is conducting is the result of a three-year collaboration agreement with Liverpool FC – the team currently leading the standings (tied with Arsenal at 71 points), including situations beyond corner kicks such as penalty kicks and predicting player movements.
TacticAI analyzes the outcomes of corner kicks with various tactics, using criteria such as which player receives the ball and whether they can shoot, to provide professional improvement suggestions. The AI’s proposals are then reviewed by five human experts, including three data scientists, a video analyst, and an assistant coach.
Measures are needed to minimize the possibility of injury
Researchers note that the experts cannot distinguish between scenarios generated by AI and real match situations, while prioritizing TacticAI’s suggestions over 90% compared to existing strategies.
Liverpool has not responded to requests for comment on whether they have implemented any changes suggested by TacticAI.
The use of data analytics to improve outcomes in sports has become increasingly complex since Michael Lewis attracted widespread attention with his book “Moneyball: The Art of Winning an Unfair Game” published in 2003. In the book, Lewis recounts how the Oakland Athletics baseball team used new player index metrics to compete with better-funded opponents.
The explosion of AI has led to potential applications of this technology in sports. The National Football League (NFL) and Amazon Web Services have developed a player health tool called “Digital Athlete” to predict and prevent potential injuries in a timely manner.