Current Agents Fail to Leverage World Model as Tool for Foresight
This identifies a critical bottleneck for agents in tasks requiring foresight, highlighting the need for better mechanisms to interact with world models.
The paper examined whether current agents can use generative world models as external simulators to anticipate future states, finding that agents rarely invoke simulation (fewer than 1%), misuse predicted rollouts (approximately 15%), and sometimes show degraded performance (up to 5%).
Agents built on vision-language models increasingly face tasks that demand anticipating future states rather than relying on short-horizon reasoning. Generative world models offer a promising remedy: agents could use them as external simulators to foresee outcomes before acting. This paper empirically examines whether current agents can leverage such world models as tools to enhance their cognition. Across diverse agentic and visual question answering tasks, we observe that some agents rarely invoke simulation (fewer than 1%), frequently misuse predicted rollouts (approximately 15%), and often exhibit inconsistent or even degraded performance (up to 5%) when simulation is available or enforced. Attribution analysis further indicates that the primary bottleneck lies in the agents' capacity to decide when to simulate, how to interpret predicted outcomes, and how to integrate foresight into downstream reasoning. These findings underscore the need for mechanisms that foster calibrated, strategic interaction with world models, paving the way toward more reliable anticipatory cognition in future agent systems.