AI table tennis robot can beat ‘elite’ players, Sony says

A team from the Japanese company built an AI-powered robotic arm clutching a ping-pong paddle that can perceive, react to and return the “the blistering speed and spin of elite-level table tennis”.

The research was published in the journal Nature.

“Ace” is a bulky industrial robot as big as the table tennis table, but is able to make split-second decisions and execute agile movements with power, according to Sony AI.

The noisy player, characterised by constant whirrs, rattles and a vigorously moving arm, faced five elite players and two professionals.

It won three out of five matches against the elite group, while other matches were “competitive,” the company said.

But Ace continued to improve its performance even after the paper was submitted to Nature, winning more recent matches against professionals with higher shot speeds, more aggressive placement closer to the table edge, and faster-paced rallies, Sony AI said.

“By solving a problem that requires exceptional real-time sensing and control, this research lays the groundwork for AI systems that can safely and reliably operate in dynamic physical environments, ranging from safety-critical settings to real-time interactive domains,” it said.

Previous table tennis playing robots could hold rallies, but never went beyond the amateur level, Sony AI said.

“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,” said Peter Stone, Chief Scientist at Sony AI.