Not the Answer, but the Ability to Choose: What 19 Years of Research Support Data Reveal About the Librarian's Evolving Role

Abstract

Seoul National University Library started research support services in 2007. Over the past nineteen years, the kinds of requests researchers bring to librarians have changed. This presentation looks at 11,424 requests from 2007 to June 2026, and asks what this change means for the role of librarians in the AI era.

The analysis focuses on the type of requests, not the total number. In earlier years, most requests were about finding previous studies and getting full text. In recent years, more requests are about decisions. Which journal should I submit to? Is my search strategy good enough? How should I read this similarity report? In the earliest period, search and access support was 94.0 percent of requests. In the most recent period, decision and workflow support was 58.9 percent.

This change matters. Researchers now have many tools, including ChatGPT, that can quickly give answers, summaries, and options. What they often need is not more information. They need to know what is enough at their stage of research. Librarians help with this. We narrow the options. We explain why we recommend something. We also share knowledge from editors, colleagues, and field conversations that is not yet in published papers. The future of research support is not about competing with AI to produce more answers. It is about helping researchers understand their options, judge what is enough, and make the final decision themselves.