Abstract
Objective – This exploratory study examines changes in traffic to academic library research guides in the early generative AI era by investigating overall traffic trends, measuring differences in usage before and after the launch of ChatGPT in November 2022, and examining changes in the referral sources that direct users to the guides.
Methods – A quantitative analysis of web traffic to the University of Ottawa Library’s LibGuides (2016-2025) was conducted to examine changes over time. Pre- and post-ChatGPT usage was compared across two equivalent 37-month periods. Google Analytics referral data (2023-2025) were analyzed to identify traffic sources and assess changes in access pathways. All analyses were conducted using RStudio.
Results – Research guide usage increased steadily from 2016 to 2022 but began declining in 2023, despite continued growth in the number of published guides. Of the 210 guides analyzed across equivalent pre‑ and post‑ChatGPT periods, most (175) experienced decreased traffic, with changes ranging from −91.3% to +462.7% and a median change of −38.35%, indicating highly variable but predominantly downward usage patterns. Although a small subset of guides (35) saw notable increases, overall engagement declined year over year. Google remained the dominant referral source, but traffic originating from Google also decreased over time, while referrals from generative AI platforms showed steady growth, alongside typical seasonal academic fluctuations.
Conclusion – Use of research guides seems to have declined since the emergence of generative AI, mirroring broader reductions in web traffic and click‑through behaviour associated with AI‑generated summaries. By empirically documenting this overall pattern within an academic library context, this study provides a necessary baseline for understanding how generative AI may be reshaping discovery pathways. Although some guides demonstrated increased engagement, the prevailing trend suggests reduced visibility and shifting access routes in an AI search environment. These findings underscore the importance of reassessing the role and discoverability of research guides for contemporary information-seeking practices.