Under Pressure: Highlights from the 2025 Library Deans and Directors Survey

Hello, and thank you for joining us. My name is Ellen Carol, and I'm a senior analyst at Ithaka S+R. >> And I'm Tracy Bergstrom. I'm the senior program manager for libraries and scholarly communication at Ithaka S+R. >> We're pleased to be here presenting highlights from our 2025 survey of library deans and directors. We conduct this work through our organization, Ithaka S+R. Ithaka S+R is part of Ithaka, a not-for-profit with a mission to improve access to knowledge and education for people around the world. >> [snorts] >> We conduct a survey of deans and directors at college and university libraries in the US every 3 years in order to track the perspectives, priorities, and strategies of these leaders over time. The most recent cycle was fielded in fall of 2025 and examined how library leaders are navigating both long-standing responsibilities and emerging challenges facing the profession. For today's presentation, we focus on four themes that together illustrate how academic libraries are navigating this moment. We'll begin by examining current hiring priorities across library areas, then turn to how libraries are responding to artificial intelligence as an emerging technology. Finally, we'll examine leaders' perspectives on data privacy before closing with how libraries are assessing and reporting their own metrics internally. Each survey cycle, we send the survey by email to library deans and directors at all four-year, not-for-profit colleges and universities in the US. We've surveyed this population since 2010, so the questionnaire includes a mix of legacy items that allow us to examine change over time, which we'll show some of those data later on, and new questions that respond to developments in the field since the prior cycle. In the 2025 cycle, we had a final sample of 483 respondents for a response rate of 36%. We have some sample demographics shown here. I do want to highlight some context on the institutions at which these library leaders work. So, the participating libraries were most commonly housed within research-intensive institutions that award doctoral degrees, and a little over half were private not-for-profit institutions. So, for our presentation, I mentioned that we focused on some specific areas of the report that we believe are most relevant to our CNI colleagues in the interest of time. But that said, the report contains a much broader set of findings ranging from how library leaders are navigating financial constraints to perceptions of open access initiatives. So, we encourage you to check out the full report if you're interested in learning more about any of these themes and the many additional findings gleaned through the survey. And with that, we will get into the results. To better understand how leaders are envisioning staffing needs in the near term and what this might signal for the workforce, we asked respondents to consider a range of functional areas and indicate whether they anticipate increasing or reducing the number of library staff in each area over the next 5 years. So, the dark blue bars show the share of respondents who anticipate staff increases, while the light blue bars show reductions. So, starting with areas of anticipated staff growth, we see a few patterns standing out quite clearly here. At the top, 31% of respondents expect to increase staff in AI and machine learning, and none anticipate reductions in this area. So, this likely reflects both the current prevalence of AI and the fact that many libraries are still building capacity here. We also see expected staff growth in more established service areas. So, 27% of respondents indicate increases in instruction or information literacy services, followed by 21% in assessment, user experience, or data analytics, and another 21% who expect to increase staffing around student success, engagement, and outreach. When we turn to anticipated staff reductions, though, the picture is less clear. Respondents were overall more hesitant to identify areas where they might reduce staffing, and there's less consensus about where those reductions might occur. But that said, a few areas of anticipated reduction do emerge. So, frontline services are most frequently identified for potential staff reductions at 19%. However, this is matched by 20% of respondents who anticipate staff increases in the same area. We see a similar pattern for technical services, where 17% anticipate reductions, but a comparable share anticipate increases. And for collection development or subject specialists, 13% staff anticipate staff reductions, um but this is again closely mirrored by those who expect growth. So, overall, while anticipated areas of staff growth in the near term, particularly around AI, instruction, and analytics are relatively clear, expectations for where staff will be reduced are often divided, with the same functional areas appearing on both sides of the ledger. But we can gain a clearer understanding of libraries' staffing trends um by looking at the areas where libraries have already downsized. So, this figure shows the areas in which staff roles were reduced among respondents who reported that their libraries had eliminated staff positions or instituted furloughs within the past 3 years. Among these libraries, 51% reduced staff in technical services, followed closely by 50% that reduced staff in frontline services. Then looking further down the chart, 33% reported reductions in collection development, subject specialists, or departmental liaison roles. Notably, the order of these past reductions closely aligns with the areas that leaders anticipated reducing on the previous slide. So, this suggests some consistency between where leaders have already absorbed staffing cuts and where they expect future reductions might occur. >> One of the major developments affecting the library space in the time between our most recent survey cycles, of course, was the proliferation of generative AI. This cycle, we included foundational questions about AI use and anticipated need for growth within libraries. To examine the perceived effects of AI in the library more closely, we asked library leaders to identify the ways they expect AI to significantly impact their library over the next 3 years. The most frequently cited impact, as we can see, is increased demand for AI literacy instruction, selected by 83% of respondents. This is followed by the integration of AI tools into library discovery systems at 74%. More than half also anticipate staffing changes or reskilling needs amongst library staff. The survey also asked the leaders where their libraries are currently employing AI tools to support internal library work or library services. Interestingly, the data indicate that libraries are integrating AI tools at different rates and in different ways. So, while there's more consensus around the ways that AI will impact the library, as we saw in the previous slide, there is unevenness around libraries' actual use of AI tools in their internal operations. The most commonly reported area of AI use was discovery and access at 40% followed by research support services and internal data analysis. We also see AI use in areas like metadata creation and validation, as well as cataloging and processing of special collections. In the previous slide, you may have noticed that 24% of respondents indicate that their library does not currently use AI tools in library operations. To further investigate this area, we asked these respondents to follow a follow-up question on what barriers, if any, were impacting their ability to use AI tools in the library. And again, we see a lot of variation across the types of barriers encountered, but those most frequently selected are issues related to strategy and staffing. Specifically, AI being a low priority relative to other library strategies was the most common barrier selected by 50% of respondents. This was followed by limited staff skills in AI at 48% and then again, ranking very closely, we have lack of time or staff capacity, as well as ethical or moral opposition to AI as commonly encountered barriers, each selected by 46% of respondents. This question, in relation to others in the survey, really calls to the forefront that strategy in relation to the implementation and and use of AI is clearly still a developing area for libraries at the present. Perhaps in response to growing needs for libraries to integrate and support AI, we can see that concerns around user data privacy are increasing. On the right, half of library leaders are concerned about third-party access to individual-level data, reaching the highest point since we first asked this question, as you can see, in 2019. In a related question we saw a few slides ago, 32% of library directors indicated they foresee heightened scrutiny of data privacy as a significant impact of artificial intelligence on the library within the next 3 years. Finally, one of the areas we're really thinking about right now is what types of data are useful to libraries to communicate their value. To this end, we introduced two questions on this topic in the last cycle. The first that you see on the screen here asks, "What types of data does your library routinely share with your direct supervisor?" Who a library director's supervisor is, whether the library organizationally reports to the provost or another entity on campus, is also tracked within the survey if you're interested in this. But here we can see that utilization data, including potentially door counts or download counts, tops the list at 77% of directors who share this information. This is followed by narrative evidence at 73% and user experience data at 66%. While responses were largely consistent across Carnegie classification and sector, leaders at baccalaureate institutions were more likely than those at other institution types to share data on teaching and learning outcomes. As a companion question, respondents were asked how frequently their library share data with their direct supervisor. 28% indicated that they share data monthly, followed closely by another 26% who report that they have no regular reporting pattern. This is an area we're continuing to follow up on, including if libraries are beginning to see changes in content use or metrics as a the of AI, and how this may affect what is reported within an institution to show engagement with the library. If you're interested in this topic, please follow up with one of us as we're just beginning this research. So, a few implications from the findings that we've shown today. First, anticipated staffing needs point to growing demand in areas such as AI and machine learning, instruction and information literacy, assessment, and student success. These latter categories have long been areas of focus for research libraries, but new demands around AI expand the library's scope of responsibility. Growing expectations around AI support are adding pressure to library staff and services. Though the scale and nature of these demands vary as we've seen. Many libraries have not yet integrated AI into their internal operations citing limited staff capacity or expertise, ethical concerns, and competing priorities. AI has also amplified long-held concerns about patron privacy as suggested by leaders increased attention to data privacy. And finally, irregular data reporting practices may limit libraries ability to demonstrate impact. This is an increasingly important challenge as leaders work to communicate value and to maintain institutional resources. We've showcased these particular issues today in this short presentation because we're using these findings to think about future work and ways in which we can support libraries. We very much value your input and reactions to these findings. Please don't hesitate to be in touch with us. Thank you so much for your attention today. The full report can be found via the QR code that you see here on your screen. We look forward to updating this group in the future about research we continue to update on these important topics. Thanks.