Missed out-of-bounds calls, betting scandals tied to player-performance props, and stars who reshape possessions without touching the ball yet vanish from the box score: the NBA is treating all three as problems technology can help solve.
The league is deploying artificial intelligence across three distinct fronts: officiating accuracy, competitive integrity surveillance, and performance analytics. Sony Hawk-Eye, a sensor-embedded Wilson basketball, and Amazon Web Services form the core of that effort, and what they do affects the game you watch.
Calling It Right: The Smart Ball and What It Changes
Optical tracking and an embedded sensor could automate the objective calls that currently send games to lengthy replay review.
Sony’s Hawk-Eye system captures player and ball movement in three dimensions with very low latency, according to an NBA partnership announcement. It supports advanced analytics and is designed to eventually automate objective decisions such as out-of-bounds rulings and goaltending calls.
Wilson is testing a basketball with a one-gram embedded sensor that measures movement, spin, and trajectory, according to Yahoo Sports. Testing has occurred in the G League, Summer League, and preseason settings, with players reportedly noticing no difference in feel.
The applications are specific. The sensor could determine which team last touched the ball before it crossed the boundary, or whether a defender’s fingertip altered the ball’s spin before a potential goaltending call.
Commissioner Adam Silver has discussed possible regular-season use as early as next season, per Yahoo Sports, but no definitive date has been set. The smart ball remains a test and a prospective officiating aid, not a system making binding calls in NBA games right now.
Draymond Green said plainly that “this is AI continuing to creep into basketball,” criticizing the direction while acknowledging that technology could reduce human error, according to Yahoo Sports. The proposed division of labor would leave referees responsible for judgment calls, particularly foul contact, while automated systems could handle measurable facts such as ball location and trajectory.
Beyond the Box Score: Integrity Monitoring and What AI Now Measures
An internal AI platform and advanced tracking tools are extending the league’s reach from the court to the betting window and deep into game context.
The NBA has built an internal AI platform that correlates basketball data, injury reports, and player-availability signals from social media, according to CNBC. It also ingests individual betting data to identify suspicious wagering patterns, with follow-up investigations and disciplinary procedures remaining human-led.
That surveillance expansion follows high-profile cases involving Jontay Porter, Terry Rozier, Damon Jones, and Chauncey Billups. The league is also tightening controls on the wager types authorized betting partners can offer, particularly player-performance “under” bets, per Yahoo Sports.
A reported player-protection initiative can detect and flag social-media harassment directed at athletes who opt into the program, according to CNBC. Deployment details and confirmed effectiveness still require further reporting.
AWS serves as the NBA’s official cloud and AI partner. Hawk-Eye tracking maps numerous body points to estimate defender proximity, shot-release conditions, and expected field-goal outcomes, context that points, rebounds, and assists never captured, per NBA Stats and AWS documentation.
“Gravity” analytics measure how much defensive attention a player draws and how that attention opens opportunities for teammates, including possessions where the star never touches the ball. A reported “Leverage Score” uses counterfactual modeling to distribute credit across players on the court, though both metrics are model outputs built on assumptions, not direct physical measurements of value.
For you watching at home, the hybrid model taking shape could mean fewer stoppages for replay reviews and more contextual data on your broadcast. The subtler challenge is accuracy in presentation: model-generated statistics carry real uncertainty, and how the league communicates that distinction will matter as much as the technology itself.
