No more poker face? AI spots players bluffing!

Artificial intelligence is finding its way into more and more areas of our lives, and the world of poker is no exception. After the World Series of Poker returned to ESPN following 2021, the channel has introduced a spectacular new feature. Omaha Productions, the company owned by NFL legend Peyton Manning and responsible for the broadcasts, incorporated artificial intelligence into its coverage.

With this ESPN is not only improving its broadcasts but also showing how technology can make poker more modern, unique, and appealing to both viewers and advertisers.

AI spots players bluffing

The system, developed by AI engineer Luke Geel, monitors subtle cues from the human body. The algorithm does not only consider players’ previous results but also analyzes behavioral patterns, such as changes in posture or the frequency of blinking. The goal is to identify moments when a player is bluffing or holding the strongest possible hand, known in poker as „the nuts”.

However, the technology does not interfere with the game itself. ESPN uses the analysis only for players who have already been eliminated from the tournament, meaning the system’s findings cannot influence ongoing hands or decisions.

Geel told Sportico that development took around six months and was much more challenging than he had originally expected. However, the project quickly attracted attention from the poker community. After presenting his work online, several players contacted him, saying they would like to use the system to study opponents or even improve their own game.

The project highlights just how important nonverbal cues are in poker. Players do not only watch the cards and chips ; they also pay attention to every detail about their opponents: the size of their bets, the speed of their decisions, their body language, or even the smallest changes in their voice.

That is why professional players spend years trying to reveal as little information as possible at the table. Sunglasses, hoodies, and other accessories have become some of the most recognizable ways players try to hide their tells.

Poker is just the beginning

The AI-powered bluff-detection technology used in broadcasts may only be the beginning. According to experts, similar systems could appear in other sports in the coming years, as more advanced technology and larger amounts of data allow AI to detect signals that are almost invisible to the human eye, such as changes in heart rate or slight changes in skin color.

These possibilities are especially interesting in soccer. For example, the research of David Freire-Obregón, a Spanish university associate professor, focuses on how artificial intelligence could help goalkeepers during penalty kicks.

According to Freire-Obregón, future AI models may be able to recognize tiny movements and body-language cues that reveal the direction of a shot before the ball is even kicked. They could also learn the unique habits of individual players, making predictions more accurate and personalized.

In 2025, Freire-Obregón’s research team investigated how reliably AI could predict the direction of a penalty before the ball was kick. The results were remarkable. By analyzing video footage of penalty kicks, a deep-learning model was able to predict with 64 percent accuracy whether the ball would go to the goalkeeper’s right or left. This was higher than the prediction accuracy of goalkeepers themselves.

However, the researcher does not believe artificial intelligence will make penalty shootouts predictable. Instead, he believes these systems could force players to hide their intentions even more carefully, creating a new chapter in the psychological battle between the shooter and the goalkeeper.

Poker as a milestone in AI development

The connection between poker and artificial intelligence is nothing new. Nearly a decade ago, in 2017, an experiment that many considered a milestone in AI development attracted worldwide attention.

During a 20-day challenge at Rivers Casino in Pittsburgh, a program called Libratus, developed by researchers at Carnegie Mellon University, faced some of the world’s best poker players. Libratus was specifically trained for no-limit heads-up Texas Hold’em, one of the most complex forms of poker. The artificial intelligence eventually achieved a convincing victory, defeating all four professional players.

However, an earlier version of Libratus had already competed against human players in 2015 but was unable to win at the time.

The improvement was remarkable. A system that had struggled against professional players only a few years earlier was now capable of beating some of the best players in the world.

The victory represented not only a technological breakthrough but also a clear sign of how quickly AI systems were advancing.