Scientists Built a Robot That Can Beat Elite Human Players at Table Tennis

Sony AI’s robot Ace combines ball tracking, spin detection, deep reinforcement learning, and an agile robotic arm to play table tennis.

In tests, Ace beat elite amateur players in three of five matches, but lost both matches against professional Japanese league players.

The result shows a robot can handle a fast, physically complex sport in real time, marking progress toward practical AI systems that interact with humans in dynamic settings.