AI agents can now check factory floors for safety without making things up

Researchers built a new way for AI agents to understand physical spaces, like factories or warehouses. This method makes sure the AI uses real measurements and facts before trying to answer questions, so it doesn't invent details.

AI systems often struggle with spatial reasoning, leading them to 'hallucinate' distances or object relationships. This new approach forces the AI to use precise calculations for spatial questions, which means it can reliably check for things like safety violations or efficient layouts in real-world settings. It makes AI more trustworthy for tasks where accuracy in physical space is critical.