New research from Brigham Young University suggests that machines that process Facebook, Google +, Twitter and email comments can outperform their human counterparts when it comes to cooperation. To demonstrate this the researchers devised an algorithm that can teach machines not only to win games, but also to cooperate and compromise.
The first waves of artificial intelligence saw computers beat humans at chess and other strategy games like Go! The second wave saw platforms out performing humans at zero sum games (mathematical representations of a situation in which each participant’s gain or loss of utility is exactly balanced by the losses or gains of the utility of the other participants).
The third wave, at least according to Brigham Young University academics, is the ability to better humans at compromise – a skill essential for the workplace and also in contracts negotiations.
According to lead researcher Dr. Jacob Crandall: “The end goal is that we understand the mathematics behind cooperation with people and what attributes artificial intelligence needs to develop social skills.”
He adds that artificial intelligence: “needs to be able to respond to us and articulate what it’s doing. It has to be able to interact with other people.”
To demonstrate this the researchers ran as study. For this thye programmed machines with an algorithm called S# and then ran them through a variety of two-player games to assess how well they would cooperate in certain relationships. The researchers tested machine-machine, human-machine and human-human interactions. In the vast majority of outcomes, machines programmed with S# outperformed humans in finding compromises that benefit both parties.
The findings have potential business applications, covering customer services and contracts assessmets. The research has been published in the journal Nature Communications and the paper is titled succinctly “Cooperating with machines.”
