Social Theory Should Be a Structural Prior for Agentic AI: A Formal Framework for Multi-Agent Social Systems

2026/05/08 by Lynnette Hui Xian Ng, Iain J. Cruickshank, Adrian Xuan Wei Lim +1 · 1 voice
Computer Science · #cs.CY #cs.MA

paper · pdf · doi:10.48550/arxiv.2605.07069

arxiv published 2026/05/08 · arxiv updated 2026/05/12

Abstract

Agentic AI systems are increasingly deployed not in isolation, but inside social environments populated by other agents and humans, such as in social media platforms, multi-agent LLM pipelines or autonomous robotics fleets. In these settings, system behavior emerges not from individual agents alone, but from the multi-agent interactions over time. Emergent dynamics of individuals in a social group have been long studied by social scientists in human contexts. This position paper argues that agentic AI systems must be modeled with social theory as a structural prior, and formalizes a Multi-Agent Social Systems (MASS) framework for how agents interact and influence to generate system-level outcomes. We represent MASS as a class of dynamical system of information generation, local influence and interaction structure, formulated by four structural priors anchored in social theory: strategic heterogeneity, networked-constrained dependence, co-evolution and distributional instability. We demonstrate the importance of each structural prior through formal propositions, and articulate a research agenda for how MASS should be modeled, evaluated and governed.

Citations

• Does Socialization Emerge in AI Agent Society? A Case Study of Moltbook
• Measuring Fine-Grained Negotiation Tactics of Humans and LLMs in Diplomacy
• Two-Faced Social Agents: Context Collapse in Role-Conditioned Large Language Models
• Evaluating Behavioral Alignment in Conflict Dialogue: A Multi-Dimensional Comparison of LLM Agents and Humans
• AuraSight: Generating Realistic Social Media Data
• AI Agents and the Law
• NomicLaw: Emergent Trust and Strategic Argumentation in LLMs During Collaborative Law-Making
• A Survey of LLM-Driven AI Agent Communication: Protocols, Security Risks, and Defense Countermeasures
• Large Language Models Miss the Multi-Agent Mark
• LLM Generated Persona is a Promise with a Catch
• Why Do Multi-Agent LLM Systems Fail?
• Advancing AI Negotiations: New Theory and Evidence from a Large-Scale Autonomous Negotiations Competition
• Building Bridges between Users and Content across Multiple Platforms during Natural Disasters
• Simulating Rumor Spreading in Social Networks using LLM Agents
• TikTok and the algorithmic transformation of social media publics: From social networks to social interest clusters
• Examining Identity Drift in Conversations of LLM Agents
• Evolution of Social Norms in LLM Agents using Natural Language
• AI models collapse when trained on recursively generated data
• The Landscape of Emerging AI Agent Architectures for Reasoning, Planning, and Tool Calling: A Survey
• Shall We Team Up: Exploring Spontaneous Cooperation of Competing LLM Agents
• Simulating Opinion Dynamics with Networks of LLM-based Agents
• SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
• Recruitment Promotion via Twitter: A Network-centric Approach of Analyzing Community Engagement Using Social Identity
• AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation
• ChatEval: Towards Better LLM-based Evaluators through Multi-Agent Debate
• AgentBench: Evaluating LLMs as Agents
• MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework
• Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena
• Improving Factuality and Reasoning in Language Models through Multiagent Debate
• Generative Agents: Interactive Simulacra of Human Behavior
• CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Society
• Neural Theory-of-Mind? On the Limits of Social Intelligence in Large LMs
• Multi-Agent Reinforcement Learning is a Sequence Modeling Problem
• Measuring Massive Multitask Language Understanding
• The spread of true and false news online
• Deep reinforcement learning from human preferences
• Attention Is All You Need
• Online Human-Bot Interactions: Detection, Estimation, and\n Characterization
• Semi-Supervised Classification with Graph Convolutional Networks
• Understanding Incentives: Mechanism Design becomes Algorithm Design
• The Role of Social Networks in Information Diffusion
• Statistical physics of social dynamics
• An overview of statistical learning theory
• Economic Action and Social Structure: The Problem of Embeddedness
• Threshold Models of Collective Behavior
• The Strength of Weak Ties
• The Role-Set: Problems in Sociological Theory

Discussions

• #toread #paper Social theory should be a structural prior for agentic AI: A formal framework for multi-Agent Social Systems by Lynnette Hui Xian Ng et al. https://arxiv.org/abs/2605.07069 [bsky, 2 points, 0 comments]