aiqbench analysis
This research exposes a critical gap between the narrative plausibility of LLM-generated agents and their ability to reproduce empirically realistic human mobility patterns in urban simulations. The study demonstrates that current models struggle with core spatial-temporal constraints like trip lengths and dwell times, often requiring explicit profile-aware initialization rather than relying on default prompts. By providing an open framework for rigorous validation against real-world data from Paris and Shanghai, the authors offer essential tools to bridge the divide between plausible storytelling and accurate behavioral simulation in AI-driven urban planning.
When Plausible Is Not Realistic: Evaluating Human Mobility in LLM-Based Urban Simulation
arXiv:2606.13835v1 Announce Type: cross Abstract: LLM-based generative agents are increasingly used in urban simulators, yet it remains unclear whether they reproduce empirically realistic human mobility patterns or merely generate plausible mobility narratives. We introduce a validation framework for evaluating the mobility of generative agents of LLM-based urban simulators against real-world mobility data. For this, we use mobility laws, temporal rhythms, network motifs, semantic activity transitions, and behavioral mobility profiles. Using datasets from the Greater Paris region and Shanghai, we evaluate AgentSociety and CitySim across multiple dimensions of mobility realism. Our analysis reveals a substantial gap between narrative plausibility and empirical mobility realism. Although the simulators capture some high-level semantic activity distributions, they struggle to reproduce core spatial and temporal constraints, including realistic trip-length distributions, origin-destination flows, dwell times, and transition dynamics. We further observe that realistic mobility diversity is unstable across default prompting configurations and may require explicit profile-aware initialization. To support reproducible evaluation, we also contribute scalable and open LLM-driven infrastructure for regional-scale map generation, observability-enhanced simulation, mobility-metric computation, and traffic simulation. Our findings highlight the need for rigorous empirical validation of LLM-based urban simulators and provide practical tools for building more realistic and reproducible urban simulation systems.
Abstract:LLM-based generative agents are increasingly used in urban simulators, yet it remains unclear whether they reproduce empirically realistic human mobility patterns or merely generate plausible mobility narratives. We introduce a validation framework for evaluating the mobility of generative agents of LLM-based urban simulators against real-world mobility data. For this, we use mobility laws, temporal rhythms, network motifs, semantic activity transitions, and behavioral mobility profiles. Using datasets from the Greater Paris region and Shanghai, we evaluate AgentSociety and CitySim across multiple dimensions of mobility realism. Our analysis reveals a substantial gap between narrative plausibility and empirical mobility realism. Although the simulators capture some high-level semantic activity distributions, they struggle to reproduce core spatial and temporal constraints, including realistic trip-length distributions, origin-destination flows, dwell times, and transition dynamics. We further observe that realistic mobility diversity is unstable across default prompting configurations and may require explicit profile-aware initialization. To support reproducible evaluation, we also contribute scalable and open LLM-driven infrastructure for regional-scale map generation, observability-enhanced simulation, mobility-metric computation, and traffic simulation. Our findings highlight the need for rigorous empirical validation of LLM-based urban simulators and provide practical tools for building more realistic and reproducible urban simulation systems.
Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
arXiv:2606.13835 [cs.CL]
(or arXiv:2606.13835v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2606.13835 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 295 words
Related coverage
Simulated Control Evaluations
LessWrong · Oct 23, 2026
VISTA: A Versatile Interactive User Simulation Toolkit for Agent Evaluation
arXiv · Jun 10, 2026
VISTA: A Versatile Interactive User Simulation Toolkit for Agent Evaluation
arXiv · Oct 14, 2026
InterviewPlayground: A Simulation Environment for Evaluating AI Interviewers
arXiv · Oct 10, 2026