OpenAI bot crushes human competition at world championship video game contest
14 Aug 2017
An artificial intelligence program just crushed its human competition at a world championship video game contest.
The outcome stunned the gaming community no end, given that bots are generally considered inferior to expert human players.
The bot from Open AI, a firm backed by serial entrepreneur Elon Musk, turned out to be a different story, and possibly a cautionary one according to commentators.
According to Open AI, its mission is to promote "responsible" AI development, or as Musk puts it, to ensure that AI did not grow unchecked to usher in doomsday of humanity.
Musk tweeted on Saturday that AI is "more [of a] risk than North Korea."
To test out a harmless use for AI, one Open AI team taught a bot to play Dota 2.
Dota 2 is an online multi-player battle game, which is something of a complex virtual version of capture the flag. Teams of players select powerful characters, called "heroes," to battle each other and the game ends when one team takes down a structure, called an "Ancient," in the opposing team's home base.
"Our bot is trained entirely through self play. It starts out completely random with no knowledge of the world, and simply plays against a copy of itself - which means it always has an evenly matched opponent," Greg Brockman, Open AI co-founder and chief technology officer said.
Musk, who founded Open AI as a nonprofit venture to prevent AI from destroying the world, had last month, told a group of US governors that AI represents a ''fundamental risk to the existence of civilization.''
According to commentators, at least for now, killer AI seems to be limited to parlour tricks at e-sports tournaments. The team at Open AI planned to extend its Dota 2 bot's capabilities into something capable of competing across the full game in a 5-vs-5 match by next year's tournament.
