Is AI More Cooperative Than Humans?

Contemporary artificial intelligence, like ChatGPT, can imitate human behaviors, and it tends to display positive traits such as cooperation, altruism, trust, and reciprocity.

A recent study from the University of Michigan employed “behavioral” Turing tests, evaluating the personality and conduct of various AI chatbots. In these tests, ChatGPT participated by responding to psychological survey questions and engaging in interactive games. The researchers then compared ChatGPT’s choices with those of 108,000 people from over 50 countries.

Greater cooperation

The researchers believe that the AI systems demonstrated a greater propensity to cooperate than humans, which may make them well suited to tasks such as negotiation and customer service.

They maintain a degree of hesitancy, however, as while we’re able to determine the output of man and machine, we’re still largely unable to ascertain “why” generative AI outputs the things it does. This distinction is likely to be crucial if we’re to trust AI systems to perform sensitive tasks, such as in healthcare.

The researchers hope that their work ultimately encourages AI researchers to work with those from other disciplines so we gain a better understanding of how AI will work alongside humans.

Facebooktwitterredditpinterestlinkedinmail