Agentic World Modeling: Foundations, Capabilities, Laws, and Beyond

2026-04-24

AI Summary

Current AI agents lack a unified and effective capability for modeling environmental dynamics, hindering their reliable performance in complex physical, digital, social, and scientific settings. This work proposes a “Hierarchy × Laws” framework that structures a three-level world model system—comprising a predictor (L1), simulator (L2), and evolver (L3)—and integrates four categories of domain-specific laws. Synthesizing over 400 studies, the project introduces a novel taxonomy to unify world modeling concepts across disciplines, establishes decision-oriented evaluation principles, and provides a reproducible benchmark suite. By surveying more than 100 systems spanning reinforcement learning, video generation, GUI/Web agents, multi-agent simulation, and AI-driven scientific discovery, the study delineates methodological approaches, failure modes, and evaluation criteria for each configuration, offering both a theoretical foundation and a practical roadmap toward building advanced agents capable of simulating and reshaping their environments.

Technology Category

• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

As AI systems move from generating text to accomplishing goals through sustained interaction, the ability to model environment dynamics becomes a central bottleneck. Agents that manipulate objects, navigate software, coordinate with others, or design experiments require predictive environment models, yet the term world model carries different meanings across research communities. We introduce a "levels x laws" taxonomy organized along two axes. The first defines three capability levels: L1 Predictor, which learns one-step local transition operators; L2 Simulator, which composes them into multi-step, action-conditioned rollouts that respect domain laws; and L3 Evolver, which autonomously revises its own model when predictions fail against new evidence. The second identifies four governing-law regimes: physical, digital, social, and scientific. These regimes determine what constraints a world model must satisfy and where it is most likely to fail. Using this framework, we synthesize over 400 works and summarize more than 100 representative systems spanning model-based reinforcement learning, video generation, web and GUI agents, multi-agent social simulation, and AI-driven scientific discovery. We analyze methods, failure modes, and evaluation practices across level-regime pairs, propose decision-centric evaluation principles and a minimal reproducible evaluation package, and outline architectural guidance, open problems, and governance challenges. The resulting roadmap connects previously isolated communities and charts a path from passive next-step prediction toward world models that can simulate, and ultimately reshape, the environments in which agents operate.

Problem

• world modeling
• environment dynamics
• AI agents
• predictive models
• governing laws

Innovation

• Agentic World Modeling
• Levels × Laws Taxonomy
• Predictive Environment Models
• Model-Based Reinforcement Learning
• Autonomous Model Revision