Knowledge base
What Are Artificial Societies?
Last updated
An artificial society is a purpose-built network of 300 to over 5,000 interconnected AI personas that represent a specific real-world audience. Unlike generic AI personas created from stereotypical demographic templates, each persona in an artificial society is constructed from real human behaviour data — how people express opinions, engage with content, and influence each other online. These personas are connected within a social graph that models influence dynamics and opinion formation, just as real human communities shape individual views. Organisations use artificial societies to predict how stakeholder groups will react to messages, strategies, or campaigns before they are deployed — generating thousands of granular, individual-level insights for high-stakes decisions.
How Are Artificial Societies Different from Standard AI Personas?
Standard AI personas are generated from broad demographic stereotypes — an LLM is told to “act like a 45-year-old male policymaker” and produces generic, predictable responses. Artificial societies take a fundamentally different approach. Each persona is constructed from real-world social behaviour data: what real people say and do online, how they engage with content, and how they react to each other. A proprietary behavioural analysis system creates an internally coherent belief system and personality for each persona, grounded in social media psychology and computational behavioural science. Critically, these personas are not isolated — they are interconnected within a social network that models how real communities shape opinions.
How Do Artificial Societies Capture Social Dynamics?
Humans are not isolated individuals — opinions form and spread through social influence. Artificial societies model this using a proprietary multi-agent orchestration system that captures social conformity forces, influence dynamics, and network effects. Rather than treating each persona as an independent respondent, the system connects personas within a social graph and simulates how people influence each other. This stakeholder network simulation is what allows artificial societies to accurately replicate the full diversity of human opinion distributions, including minority viewpoints and emergent consensus patterns that individual-level approaches miss entirely.
What Audiences Can Artificial Societies Represent?
Artificial societies can represent any audience whose members can be observed online: consumers in a category, employees in a business unit, investors, journalists, policymakers, regulators, clinicians, or the residents of a single city. Each society is built to the audience the decision actually depends on, from a few hundred personas for a niche professional group to several thousand for a national population.
Is Artificial Societies the Same as the Academic Term?
Artificial Societies is an AI audience-simulation company founded in October 2024 and headquartered in London. It is distinct from the academic term “artificial societies” (agent-based social simulation, Epstein & Axtell, 1996): the company builds simulation software that models how real audiences form, spread and change opinions, so organisations can test consequential decisions before making them.