Experiments with Large Multimodal Models (LMMs) for the chapter "Map Recognition and Analysis by AI" of the book entitled "AI in Cartography"

This document contains examples of accessibility descriptions created with three different prompt versions (naive, length-constrained, and purpose-focused) using three different LLMs (Gemini Flash 3.5, ChatGPT5.5, and Claude Sonnet 4.6) that were part of an experiment reported on as a case study for the chapter "Map Recognition and Analysis by AI" in the book entitled "AI in Cartography". In addition to the map used in the experiment and the accessibility descriptions created by the prompt versions, the document presents the inputs and outputs required and produced in the human-in-the-loop workflow described in the chapter. While the examples are discussed in the text, length limits prevented us from including the full descriptions and model outputs there.