An AI ping pong robot is beating top human players, but don’t panic just yet

If you tend to fear AI-driven robots replacing human workers in physically demanding jobs, consider this your first warning.

A robotic arm developed by Sony, also called Ace, was recently called “the first autonomous system to compete with elite human table tennis players.” That’s a quote from a study spread on the front page of The environmentthe world’s most respected peer-reviewed scientific journal.

Ace researchers have brought receipts. As you can see in the video above, the eight-joint robotic arm is able to make split-second decisions with AI fed real-time data from nine cameras. Scored a lot of points and won a few games against the world’s top ping-pong players at Sony HQ in Tokyo.

But here’s the good news buried in all the data. Yes, within the confines of this study, Ace was competitive. That doesn’t mean Ace can figure out how to win every time; it’s not like a robot running a half marathon that has to be able to run at one speed. And, worst of all, human players begin to see the flaws in Ace’s ping-pong strategy.

Ace is not the first robot to play ping-pong. Researchers have long been interested in the game because of its speed and real-time decision-making, which is a major frontier in robotics. In this case, the Ace marks the milestone of the AI ​​system and the most reliable arm.

That arm was able to track a ping pong ball with 10 milliseconds of latency – more than 10 times faster than the human brain can handle.

“Ace’s cutting edge skills are fully trained in simulation using reinforcement learning, and transferred directly to a real robot,” Sony explained in a blog post. “This is like a player who practices endlessly in the training hall and then enters the real field without having to relearn anything.”

But that’s just it – ping-pong players learn on the go, and they look at more than just the ball.

Mayuka Taira, who lost to Ace last December, told Sony that the robot scared her at first. “Because you can’t read its reaction, you can’t feel what kind of gun it doesn’t like or fight against, and that makes it very difficult to play against it,” he said.

But Rui Takenaka, who lost and won to Ace, went to that important human step. Here’s what the company had to say, emphasizing ours:

When I used a serve with complex spin, Ace also returned the ball with complex spin, which made it difficult for me. Butwhen I used a simple server, what we call a knuckle serve, Ace returned a simple ball. That made it easier for me to attack with the third shot, and I think that was the main reason why I was able to win.

Did you get that? Ace, the most intelligent system, bit the knuckle service.

“Professional athletes are very good at adapting to their opponents and finding weaknesses, which is where we work,” Ace project leader Peter Dürr told Reuters.

So we shouldn’t put away our ping pong rackets just yet. But we should be more concerned about the mention of security claims in various reports and blogs about Ace.

Because the most profitable real-world applications of fast systems like these don’t exist in the Olympics. It’s on the battlefield — where going faster than the human eye can mean game over for human soldiers.