Generative Artificial Intelligence (GenAI) tools have experienced explosive popularity since OpenAI made ChatGPT publicly available in November 2022. The tool is based on a generic Large Language Model (LLM) wrapped in a dialog system that enables it to probabilistically generate meaningful answers to user prompts. In these short two-and-a-half years, OpenAI has released several newer model versions and ever more powerful tools, while numerous other vendors started offering their own products.

In this dynamic environment, it is surprisingly difficult to identify an authoritative list of GenAI capabilities. Tentatively, these include (but are not limited to) creation of texts and other media; conversation (chat); creation of test datasets; code generation and execution; style modification, translation, and personalization of texts; summarization of materials; extraction of structured data; classification of information; task planning; as well as web search and synthesis. Some of these functions are available only in the more advanced, subscription-based tools, which are beginning to compete with human research assistants.

In this presentation, I will share a small number of experiments using different models within ChatGPT Plus for purposes related to cartographic research, teaching, and practice. These include data extraction from documents; table join; calculating descriptive statistics; chart and map creation; web map coding; as well as map interpretation and assessment. Owing to the pace and lack of transparency in GenAI development, academic researchers seem to be playing catch-up with the latest technology. A systematic account of GenAI applications in Cartography could help focus development efforts as well as guide critical assessments of the benefits and risks of this technology in our field.