Published: January 20, 2019

Is AI better than punters at bluffing in gambling?

Gambling has partly made a digital shift, away from dimly-lit casinos to the brighter online world, opening up opportunities for more people to lose their money (and a few to win). What role can, and should, AI and machine learning play?

It follows that as gaming and gambling become digital activities so the use by gamblers of modern computing techniques, like artificial intelligence and machine learning, follows. While gaming companies will wish to use such technologies to minimize their losses and to maximize their profitability, gamblers will also be keen to assess how such technologies can help them to increase their chances of beating the house and coming away winning.

What is certain is that on an even-base artificial intelligence is most likely to beat a human at many cerebral pursuits. For example, in 2017 Digital Journal reported how Google's DeepMind AlphaGo artificial intelligence platform twice defeated the world's number one Go player Ke Jie.

Machines too are vulnerable. Another invention from Google - AlphaZero - defeated what was said to be the world's best chess program, Stockfish 8, after studying the complexities of chess for just four hours. In a series of contests, AlphaZero won or drew all 100 games played.

But what about the human powers of prediction and ingenuity that come into play with gambling? How does the hot-shot 'professional' gambler fare against artificial intelligence? According to Engadget advanced forms of artificial intelligence, such as Libratus system from Carnegie Mellon University, are able to match and beat human gamblers.

Commenting on the ingenuity of machine intelligence, Libratus co-developer Noam Brown said: "People think that bluffing is very human -- it turns out that's not true. A computer can learn from experience that if it has a weak hand and it bluffs, it can make more money."

From the alternative standpoint — artificial intelligence helping gamblers — Auckland University of Technology and the Nelson Marlborough Institute of Technology researchers showed how machine learning applied to sport result prediction is capable of estimating the outcome of sports with a high degree of success.

© Public Gaming Research Institute. All rights reserved.