How do people understand maps and will AI ever understand them?

<p dir="ltr">The human brain is an incredible piece of cognitive machinery. Maps are similarly incredible. Although not every map user finds understanding a map to be an easy task, like every tool that humans use, with practice, it is possible for most people to learn to understand and use maps effectively. Maps encode a large amount of knowledge and can be useful in supporting a wide variety of activities, both professional and personal. But what does the future hold for maps? In this research, we explore what we know about one rapidly evolving technology, artificial intelligence (AI), and what it might mean for how maps (do) work.</p><p dir="ltr">Recent news articles with titles such as ‘<i>Heard AI is coming for your job? For these copywriters, that ‘future’ arrived months ago</i>’ have explored whether general-purpose AI tools, such as chatbots, can perform tasks traditionally carried out by people. Although cartographers have been experimenting with AI for making maps for several decades and researchers continue to develop new AI-supported approaches for making maps (see Kang et al., <a href="https://www.tandfonline.com/doi/full/10.1080/23729333.2025.2481692#" target="_blank">Citation2024</a> for a recent review), we will address a different question here: How might AI tools affect how we use maps? This, in turn, prompts the related, more fundamental question of whether AI tools can understand maps.</p><p dir="ltr">To explore these questions we must first know what it means to understand a map. This depends on how we conceptualize maps. Is a map a designed artifact through which its maker aims to communicate a specific message? Is it a collection of arguments or propositions about the world that enable action to be taken? Is it a storehouse of information that is ready to be mined for wisdom? Or is it an aesthetic object (perhaps even an artistic object) that can be evocative, communicative, or both? We focus here on maps as informational and analytical graphics, recognizing that understanding maps as artistic and asethetic objects deserves its own fulsome treatment. We begin by looking at a selection of existing cartographic theoretical frameworks to examine what it means to understand a map. We can then use these ideas to assess to what extent today’s AI tools can understand maps and explore how their ability to do so might develop as AI tools continue to evolve in the future.</p>