WSOP's AI tells detector debut sparks poker debate, pros skeptical

At the 2026 World Series of Poker Main Event, ESPN introduced a flashy new AI-powered "tells detection" tool that overlays live metrics during broadcasts. Designed by Luke Geel, an AI engineer for the US Air Force, the system watches every hand captured on the tournament's three camera tables and builds a database of player behaviors—eye movements, blink rate, posture, chip handling, and hand fidgets. It correlates these inputs with hand outcomes to estimate the likelihood that a player holds a strong made hand, a drawing hand, or a bluff. The tool periodically appeared during early July broadcasts, and the slick graphics immediately divided the poker community.

Professional players remain deeply skeptical. The biggest issue is data scarcity: over 9,000 entries competed in the Main Event, but the vast majority never sat at a camera table long enough for the AI to build a robust personal profile. Michael Gagliano, a 17-year pro who made this year's final table, combed through every second of ESPN's streams during the break but found the limited screen time made it hard to extract reliable tells—for humans or AI. Shaun Deeb, a two-time WSOP Player of the Year, noted that real-world tells are more abstract than Hollywood's famous "Oreo cookie" scene from Rounders. While the tool looks impressive, pros question whether it truly captures the nuanced, context-dependent nature of human behavior at the felt.