Artificial Intelligence Wins $800,000 Against Human Poker Pros

artificial intelligence poker AI
Cards during the World Series of Poker Main Event at the Rio Hotel & Casino, Las Vegas, October 31, 2012. REUTERS/Las Vegas Sun/Steve Marcus

The latest battle between man and machine is being played out on the poker tables of a Pennsylvania casino, and so far humanity is losing.

An artificial intelligence computer program called Liberatus has accumulated winnings of almost $800,000 against a team of professional poker players at the Brain Vs. Artificial Intelligence competition at Rivers Casino in Pittsburgh.

Liberatus, developed by Carnegie Mellon School of Computer Science, is aiming to be the first computer program to win a professional poker tournament—a game considered by many AI researchers to be one of the hardest for computers to beat humans at.

Artificial intelligence advances in recent years have seen computers master and eventually outperform the best human players at games including Chess and the game show Jeopardy!.

In 2016, researchers at Google's DeepMind developed a system that was able to beat champion Go player Lee Sedol—a significant step for artificial intelligence. But a victory for Liberatus could be considered an even greater milestone.

'Human' skills like bluffing are required to successfully compete at poker, as well as the randomizing of actions in order to confuse opponents.

A predecessor to the Liberatus program lost in a similar No-Limit Texas Hold'em tournament in 2015, leading several of the professional poker players to underestimate the bot's ability.

"The bot gets better and better every day. It's like a tougher version of us," poker player Jimmy Chou said.

"The first couple of days, we had high hopes. But every time we find a weakness, it learns from us and the weakness disappears the next day."