The milestone victory of AlphaGo, an artificial intelligence system developed by Google, in 2016 has spurred a Go renaissance that has led to human players becoming more creative in their game strategies, according to a recent study published in the journal PNAS. Researchers from the City University of Hong Kong and Yale analyzed a dataset of over 5.8 million Go moves made during professional play between 1950 and 2021 with the help of a “superhuman” Go AI program that grades the quality of moves. They found that the quality of professional play improved relatively little before 2016, but since the rise of superhuman AIs in 2018, median decision quality index (DQI) values have changed at a rate above 0.7.
The study revealed that players have employed more novel strategies in recent years, with 88 percent of games in 2018 featuring a combination of plays that had not been observed before, compared to 63 percent in 2015. The researchers suggested that the development of superhuman AI programs may have prompted human players to explore novel moves and break away from traditional strategies, which in turn may have improved their decision-making.
The study’s findings highlight the significant impact that artificial intelligence has had on traditional games like Go, which were previously thought to be dominated by human players. The victory of an amateur Go player against one of the game’s top-ranked AI systems earlier this year, using a strategy developed with the help of a program designed to probe AI systems for weaknesses, further underscores the growing sophistication of human players in the face of AI advancements.
