Developing Library Consultation Methods to Assist Students Who Use Large Language Models in the GIS Field

As students become more familiar with large language models (LLMs) and the programs that use them, such as ChatGPT or Copilot, they may begin incorporating these tools into their research and schoolwork. This implementation introduces a new element into library consultations for GIS and map librarians, as the inclusion of LLMs brings forth benefits, like increased efficiency, as well as complications, like output inaccuracy, into these already varied discussions that can range from questions on locating geospatial resources to complex technical queries about specific GIS software. The problematic aspects of these tools, as well as issues related to user error, provide insight into what GIS and map librarians need to focus on to adequately prepare for consultations involving this technology, as well as how to approach students who use it. Drawing on existing literature about LLMs, geographic information systems (GIS), and librarianship, this paper suggests employing library consultation methods focused on general AI literacy instruction for responsible and critical LLM use, assistance with effective prompt writing techniques, education in mitigating model issues, and, where appropriate, suggesting alternatives to LLMs to better prepare students to engage with these tools for their own geospatial work.