Researchers at Hugging Face have published a paper proposing a systematic blueprint for building Economic World Models (EWMs) — generative AI systems that simulate how economies evolve from within by modeling heterogeneous agents, their beliefs, actions, and the market and institutional mechanisms that shape aggregate outcomes.
The paper organizes EWM systems into a six-level capability ladder, ranging from fixed rule-based agent worlds to adaptive and LLM-based agent worlds, self-evolving agents, evolving institutional worlds, and finally sim-to-real economic twins aligned with real-world observations.
A systematic literature survey across these levels reveals that existing research is concentrated in lower-level agent and simulation environments. Systems featuring self-evolving agents, endogenous institutions, persistent empirical alignment, and validated economic mechanisms remain rare.
The authors position EWMs as high-fidelity sandboxes for human decision-makers and as training, planning, evaluation, and safety substrates for AI agents. They also release a curated paper list and related resources to support future research in this emerging field.