Social World Models

2025-08-30

AI Summary

AI struggles to intuitively model implicit context and dynamic evolution in social interactions. This paper introduces S3AP—the first structured representation framework for social reasoning—formalizing the social world as a quadruple of state, observation, action, and mental model, and pioneering the application of partially observable Markov decision processes (POMDPs) to unify theory-of-mind reasoning and future situation forecasting. Methodologically, it leverages large language models to parse social states from natural language and trains an end-to-end social world model to support dynamic prediction and decision optimization. Evaluated on five benchmark tasks, S3AP achieves substantial improvements: +51% accuracy on FANToM theory-of-mind reasoning and +18% performance on SOTOPIA interactive decision-making—establishing new state-of-the-art results across all tasks.

Technology Category

• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Knowledge Representation and Reasoning: Reasoning with Beliefs

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Humans intuitively navigate social interactions by simulating unspoken dynamics and reasoning about others' perspectives, even with limited information. In contrast, AI systems struggle to automatically structure and reason about these implicit social contexts. In this paper, we introduce a novel structured social world representation formalism (S3AP), designed to help AI systems reason more effectively about social dynamics. Following a POMDP-driven design, S3AP represents social interactions as structured tuples, such as state, observation, agent actions, and mental states, which can be automatically induced from free-form narratives or other inputs. We first show S3AP can help LLMs better understand social narratives across 5 social reasoning tasks (e.g., +51% improvement on FANToM's theory-of-mind reasoning with OpenAI's o1), reaching new state-of-the-art (SOTA) performance. We then induce social world models from these structured representations, demonstrating their ability to predict future social dynamics and improve agent decision-making, yielding up to +18% improvement on the SOTOPIA social interaction benchmark. Our findings highlight the promise of S3AP as a powerful, general-purpose representation for social world states, enabling the development of more socially-aware systems that better navigate social interactions.

Problem

• AI systems struggle with implicit social context reasoning
• S3AP formalism structures social dynamics for better AI understanding
• Enables prediction of social interactions and improves agent decision-making

Innovation

• S3AP formalism for structured social representation
• POMDP-driven design with structured tuples
• Inducing social world models from narratives