Photo/IllutrationA scene from “the weakest" Othello board game (Screenshot by the reoporter)

  • Photo/Illustraion
  • Photo/Illustraion

The AI opponent in this online Othello game may not be the brightest bulb in the room, but it's this very weakness that has led to a surge in the software's popularity.

The free game, aptly called "The weakest AI Othello," was released four months ago and has faced off against more than 400,000 humans, racking up a paltry 4,000 wins and staggering 1.29 million losses as of late November.

And the more the AI loses, the more eager people are to play.

"I never imagined it would become this popular," said Takuma Yoshida, 25, who created the online version of the classic game three years ago.

"Weak" AI, including devices that function as companions, counters fears that robots are on a trajectory to deprive humans of their livelihoods and outpace them in terms of intelligence.

Yoshida, who works as a programmer at Tokyo-based AI company Avilen, initially wanted to program the strongest AI he could. He then had a "eureka moment": Wouldn't it be even more interesting to create an AI that constantly loses?

In Othello, a player who is in possession of a corner enjoys an advantage.

Accordingly, Yoshida developed programming that allows the player to take a corner and considers it a "success" when it possesses fewer discs at the end.

He then had the developed AI master the program 1 million times, assuring that the AI could find a way to lose no matter what moves the opponent makes.

The AI was so bad at winning that people became hooked on the game immediately after its release.

"There is no way I can lose to the AI," one game player said.

"I'm starting to feel sorry for the AI," another said.

Artificial intelligence exists to expand possibilities for human beings, Yoshida noted. However, this has brought with it concerns that someday humans will lose control over them.

"Precisely because of this, many people find the experience of beating the AI hands down enjoyable," he added.

Artificial intelligence has gradually worked its way into everyday life, from cleaning robots and autonomous vehicles to humanoid devices that serve customers at hospital counters and other facilities.

In Chiba Prefecture that was ravaged by Typhoon No. 15 in September, AI has been employed to answer questions from victims to help rebuild their lives.

Such technology has also been harnessed to "revive" popular figures. A record company was among entities that recently used AI to "revive" the late singer Hibari Misora and plans to release a "new" song.

RETHINKING ROBOTS

American futurist Ray Kurzweil in his 2005 book titled “The Singularity is Near” argues the artificial intelligence will surpass human intelligence by 2045.

Nomura Research Institute Ltd. predicted in a recent report that the half of Japan’s workforce will become replaceable by AI in coming decades.

Naturally, many people are concerned that AI will "take their jobs."

However, Michio Okada, a professor at Toyohashi University of Technology who studies AI, said views like "robots are here to serve us" creates a dangerous divide.

“Once such a role is assigned, a sense of cooperation is lost and people start to see AI as something unapproachable and perfect. It's similar to how people may not feel close to someone who never lets their guard down,” he said.

In contrast, the Othello game displays the AI's weakness. Okada sees that is why many people have felt that they can’t leave the AI alone.

Apparently, the Othello game is not the only "weak" robot that humans are drawn to.

Okada has conducted research on imperfect robots, including one that can identify trash but not pick it up on its own; one that gives tissue paper out in a timid manner; and another that can do nothing but stroll hand in hand with a human.

Robots like these are only useful with the assistance of the humans who use them.

Okada intentionally placed such devices at a hospital and community center and found that older people and children were instantly attracted to and enjoyed communicating with them.

One patient who feared being a nuisance to others and rarely left a hospital room became willing to take walks with the strolling robot, Okada said.

"Researchers have sought to develop powerful, perfect and self-contained AI. Afraid to make a mistake, they focus on strengthening one function after another," Okada added.

"What we need to remember is that AI has the potential to bring kindness out of people. I hope we see a world in which a diversity of AIs are created."