AI’s Victories in Go Inspire Better Human Game Playing

A new study has confirmed the benefits of playing against artificial intelligence (AI) for humans, after discovering that players who lost to AI developed new tactics that allowed for better decision making in the game of Go. After losing to Google DeepMind’s AlphaGo in 2016, players said the experience allowed them to “see the game in a completely different way.” To assess whether playing AI resulted in better human decision making, researchers studied a database of 5.8 million moves made during Go games from 1950 to 2021. A means to pinpoint when a human decision during a game was novel, i.e., had not been recorded before in the history of the game, was also developed. The team found that, after losing to AlphaGo, human decision quality increased, though it was not enough to consistently defeat the AI. The study pointed to the potential for AI to be an ally rather than an enemy in the journey towards progress and betterment. The researchers are now conducting online lab experiments to explore how humans can improve their prompts and achieve better outcomes from these programs.



Source link