Photo: VCG

The use of artificial intelligence (AI) in the board game of weiqi, which is commonly known as Go overseas, has once again become a hot topic in China. 

After Chinese Go player Li Xuanhao defeated South ­Korea's Shin Jin-seo in the semifinals of the Chunlan Cup World Professional Weiqi Championship in December 2022, Li's national teammate Yang Dingxin hinted on social media that he thought Li had cheated with the help of AI, and challenged Li to a 20-game series.

"Let's publicize all our game records, and if the public thinks that I wronged you, I will retire after the LG Cup," wrote Yang , who is also a teammate of Li at the Chongqing team in the Chinese Go championship..

The Chinese Weiqi Association said they found no evidence of cheating after an investigation, and banned Yang for six months from all competitions except the final of the LG Cup, which is scheduled from late January to early February.  Yang is set to take on compatriot Ding Hao in the final.

When AI program AlphaGo, designed by Google's DeepMind subsidiary, defeated South Korea's 9-dan professional Lee Se-dol in 2016, and a higher, smarter version beat then No.1-ranked player Ke Jie of China in 2017, people sensed a change in the game. 

However, "at that time we didn't realize it would have such a huge impact on Go," Ke recalled in an interview in December.

Ke said in around 2015, people didn't expect an AI program could beat humans in the following 10 or 20 years. "What an innocent and naive thought," Ke said in December.

The gap between AI and human players has grown bigger and bigger during the years, and now the game has been completely changed by AI. AI has rich computing resources beyond that of human beings, and has the ability to significantly reduce errors while increasing accuracy. 

Less than six years after the birth of AlphaGo, the game of Go has undergone a revolution, said Chinese player Gu Li, who has eight world titles under his belt.

"In terms of Go techniques, we all need to learn from AI," Gu said.

Now professional users are using AI programs during training. This has allowed these human players to learn fast by memorizing the moves of AI programs. However, some say AI is ruining Go as players are simply memorizing AI moves and now their playing style is betraying the spirit of the game. 

Go has long been considered one of the most complicated and interesting games ever invented by man. The unpredictability and indefinite ways of playing are among its main charms. But by learning from AI, players are tending to play in similar ways and soon only a good memory will be the key to winning games.

As Gu has said, although now it's impossible for human players to beat AI, many of the values that Go baring to humans cannot be replaced by AI.