Sony's ping-pong Ace robot stuns the world, takes down top human champions

If you want to prove capability, why not compete? AI companies are doing just that—through games, and Sony is no different. Its Ace robot, in table tennis matches, beat even some expert ping-pong players. Robots powered by AI are increasingly participating in competitions such as running and marathons to prove the skills required to interact in real time with humans.

On April 20, a Chinese humanoid robot completed a half marathon against human racers, winning the competition with a time about seven minutes faster than the men’s world record. Before this, Chinese robotics startup Unitree Robotics showcased its H1 humanoid robot reaching a sprint speed of up to 10 metres per second, marking a world record for a humanoid robot.

And now, Japanese company Sony has showcased the speed and precision of its technology through a robot dubbed “Ace.”

Sony’s Ace won seven out of 13 games it played against five different elite athletes, defined as those with more than 10 years of intensive training. It won one out of seven games it played against two professional players who regularly compete in professional leagues.

A test of precision and reaction

While the news is amusing—that a robot is able to compete with humans in games such as table tennis, which demands precise movements, split-second decisions, and powerful reactions—it also shows how far the technology has reached in understanding complex physical situations.

Previously, robots have been able to hit a ping-pong ball back and forth, but none have surpassed an amateur level. Ace matched the speed and responsiveness of the human mind and body in a game where the ball moves at high speed and spin can alter its trajectory.

Inside the technology powering Ace

Ace’s performance also shows Sony’s advanced technology. It was Sony that ultimately powered the robot with eight joints to control paddle position, adjust overall orientation, and deliver powerful shots.

These motors were assisted by a complicated vision system made up of nine traditional cameras. These cameras assisted Ace in locating the ball in space, and a gaze control system helped it measure the ball's angular velocity and spin to help Ace adjust its trajectory accurately.

A step forward for real-world AI

The new research outlining Ace’s capabilities and achievements was published in Nature and marks the first time a robot has been competitive with experienced players.

The research shows “that an AI system can perceive, reason, and act effectively in complex, rapidly changing real-world environments that demand precision and speed,” said Peter Stone, chief scientist at Sony AI, in a statement to Bloomberg. “Once AI can operate at an expert human level under these conditions, it opens the door to an entirely new class of real-world applications that were previously out of reach,” he added.