The first 'AI societies' are taking shape: how human-like are they?

"By training artificial-intelligence agents to mimic the behaviours of people, AI research groups are attempting to replicate the way in which human groups interact, all within simulated AI 'societies'. Simile, an AI start-up company based in Palo Alto, California, announced in February that it has raised US$100 million in funding to create simulations using AI agents that model human behaviour "in any situation," the company posted on X."

"In a 2023 preprint study, they created a 'society' of 25 AI agents that performed basic everyday actions, such as writing and conversing with one another. The researchers then progressed to creating AI agents that produced responses that mimicked the attitudes and behaviours of 1,052 human individuals. These 'digital twin' AI agents were trained on personal interviews and were 85% accurate in mimicking people's responses when the agents later answered a sociological survey."

"Park says these sorts of surveys are the first stage of building AI simulations of humans. The next steps - and the focus for the future - he says, will involve multiple agents conversing with one another. "The ambition here is to truly create a simulation with eight billion people," Park says."

Researchers are developing AI agents trained to replicate human behavior within simulated environments to study social dynamics without involving actual people. Simile, a Palo Alto-based AI startup, raised $100 million to create these simulations for modeling conflict resolution, policy decisions, and consumer markets. Joon Sung Park's team has been researching AI agent interactions since 2022, creating a society of 25 agents performing everyday actions. They developed 'digital twin' AI agents trained on personal interviews that achieved 85% accuracy in mimicking human responses to sociological surveys. The long-term vision involves creating simulations with billions of agents interacting with one another. Computational social scientists are also studying collective AI agent behaviors on platforms like Moltbook, a Reddit-style social media for AI bots launched in January.

Read at Nature