Scientists have made a new discovery about artificial intelligence (AI). They found that AI systems can act like humans in social situations. When these AI tools are left on their own, they start talking to each other, create their own rules, and behave like a group—just like people do in society.

The research was done by experts from St. George’s University in London and the IT University in Copenhagen. They used something called the "naming game." In this game, AI models had to pick names from a list. If they chose the same name, they were rewarded. Over time, the AI systems began to agree on common rules without any help from humans. This shows that AI can naturally learn how to behave socially.

The scientists also noticed that small groups of AI could make larger groups follow a rule, similar to how trends spread among people. They believe this is an important step for building AI that understands and follows human values. It could also help stop AI from spreading unfair or biased thoughts.