Sony’s groundbreaking robotic table tennis player has successfully challenged and occasionally defeated professional human competitors, marking what researchers describe as a historic breakthrough in artificial intelligence and robotics capabilities.
The Japanese technology company developed the mechanical athlete, dubbed Ace, and tested it against skilled human players. The robot demonstrates remarkable abilities through its nine strategically placed cameras surrounding the playing area and its exceptional capacity to track the ball’s logo for spin analysis.
Ace mastered the sport through reinforcement learning, an AI training technique that allows machines to improve through practice and experience.
“There’s no way to program a robot by hand to play table tennis. You have to learn how to play from experience,” explained Sony AI researcher Peter Dürr, who co-authored the research published Wednesday in Nature journal.
Sony constructed a regulation-sized table tennis facility at its Tokyo headquarters to ensure fair competition between the robot and highly trained human athletes, Dürr told The Associated Press. Professional players expressed amazement at Ace’s skill level during testing.
According to Sony, this represents the “first time a robot has achieved human, expert-level play in a commonly played competitive sport in the physical world — a longstanding milestone for AI and robotics research.”
The specially designed robot features eight articulated joints that control its movement range, allowing it to position its paddle, deliver shots, and quickly counter opponent strategies.
“Speed is really one of the fundamental issues in robotics today, especially in scenarios or environments that are not fixed,” stated Michael Spranger, Sony AI’s president, during an interview.
“We see a lot of robots that are in factories that are very, very fast,” Spranger continued. “But they’re doing the same trajectory over and over again. With this technology, we show that it’s actually possible to train robots to be very adaptive and competitive and fast in uncertain environments that constantly change.”
Spranger indicated such advances could benefit manufacturing sectors and other industries, though he acknowledged the potential military applications of such rapid, perceptive technology.
While a humanoid robot recently outpaced human marathon records in Beijing, creating a machine capable of real-time interaction and competition with skilled human athletes presents unique technical challenges.
Spranger emphasized the importance of maintaining competitive balance, ensuring the robot’s speed, reach, and performance matched those of dedicated athletes who train at least 20 hours weekly. Ace operates under official table tennis regulations on standard court dimensions.
“It’s very easy to build a superhuman table tennis robot,” Spranger noted. “You build a machine that sucks in the ball and shoots it out much faster than a human can return it. But that’s not the goal here. The goal is to have some level of comparability, some level of fairness to the human, and win really at the level of AI and the level of decision-making and tactics and, to some extent, skill.”
This approach means “the robot cannot just win by hitting the ball faster than any human ever could, but it has to win by actually playing the game,” he added.
Computer scientists have traditionally used board games like chess to measure AI progress, later expanding into complex video game environments. However, transitioning AI from digital simulations to physical world applications remains the ultimate goal for robotics developers.
The previous year represented a “kind of ChatGPT moment for robotics,” according to Spranger, with innovative AI methods emerging to help robots understand real-world settings and perform physically challenging tasks, including acrobatic maneuvers.
Sony joins other technology companies exploring robotic table tennis. John Billingsley pioneered such research in 1983 with his paper “Robot Ping-Pong,” while Google’s DeepMind division has also investigated the sport recently.
Despite the achievement’s significance, Billingsley suggested Sony’s comprehensive computer vision and motion tracking systems create unfair advantages against human opponents limited to two eyes.
“I would not want to belittle the achievement, but they have gone at the task mob-handed, and used sledgehammer techniques,” wrote Billingsley, a retired mechatronics professor from Australia’s University of Southern Queensland, in his AP correspondence.
Nevertheless, he acknowledged that the work reinforces how “true progress comes out of contests, whether they involve hitting a ball or setting foot on Mars.”
Japanese professionals Minami Ando and Kakeru Sone participated in competitions against Sony’s robot, with Japanese Table Tennis Association officials serving as match judges.
Following the Nature publication submission, Sony researchers continued development, reporting that Ace increased shot velocity and rally intensity while adopting more aggressive play closer to table edges. In December testing against four skilled competitors, Ace defeated three of the four players.
Veteran player Kinjiro Nakamura, who represented Japan in the 1992 Barcelona Olympics, observed Ace execute a particularly impressive shot and remarked that “no one else would have been able to do that. I didn’t think it was possible.”
However, witnessing the robot’s success “means that there is a possibility that a human could do it too,” Nakamura concluded in his published comments within the Nature study.
