Abstract
This special issue of the IFLA Journal makes a wonderful contribution to the library and information science (LIS) community's understanding of the impact of artificial intelligence (AI) in libraries, across sectors and continents. Eighteen varied articles reflect the multifarious impacts and uses of AI, and constitute an important collective addition to our practical and scholarly knowledge about AI in our context. Inevitably, this special issue does not offer definitive answers or a comprehensive picture: that is scarcely possible, not least because the technology is rapidly changing, and given the global scope of libraries. But the special issue does bring together fascinating new evidence, thought-provoking insights on current debates, and useful resources and reference points. Combined with other recent IFLA publications, such as Balnaves et al. (2025), this special issue reveals the vibrancy of the profession's response to AI.
Broadly speaking, the articles could be said to fall into five categories, which are strongly interlinked. First, we have two articles exploring the debate defining AI literacy in the context of information literacy and digital literacies (Mercer et al., 2025; Ali and Richardson, 2025). Second, there are three studies investigating AI use and measuring AI literacy among specific groups or in specific geographical contexts (Subaveerapandyian et al., 2025; Bui, Do and Dinh, 2025; Hossain et al., 2025). A third category of articles relates to library use: How are libraries using AI in services and do they have a strategy or policy? There is also an article prompting us to reflect on the link between library and wider institutional policies (Martínez-Camacho et al., 2025; Ngulube and Molaudzi, 2025; Hamad and Shehata, 2025; Buitrago-Ciro et al., 2025; Tella et al., 2025; Wilson, 2025). A fourth group of articles is about the evaluation of third-party tools, including AI chatbots, for information uses in particular contexts or for specific tasks (Simmons et al., 2025; Elsawy et al., 2025; Wang et al., 2025; Rao et al., 2025). There is also an article presenting a valuable general model of criteria to evaluate such tools (Nogueira et al., 2025). The fifth and final group of articles deals with the development of AI-based services by libraries (Cifuentes-Silva et al., 2025; Jhan et al., 2025). These articles are a little more technical and less generative-AI-focused. They are potentially some of the most important articles because we gain a sense of real, practical uses of AI, often developed over considerable time to meet known user needs.
Reflecting on the findings of the articles in this special issue, we can discern several patterns. The majority of the articles analyse the impact and use of generative AI, especially AI chatbots, such as ChatGPT. At least at present, other forms of AI, such as for describing content or prediction, seem to have faded into the background of attention.
Indeed, it seems that AI chatbots are having a global impact, with our special-issue articles revealing not only some familiar patterns (such as the popularity of chatbots among students) but also some markedly differential implications. The focus of interest of the potential uses of AI in education and in health contexts is a bit different, for example (Simmons et al., 2025). We have evidence of significant geographical variation, such as between levels of AI literacy among LIS students (Hossain et al., 2025). In fact, the evidence presented in the special issue suggests that the use of AI in library services is still in its infancy in many regions (Buitrago-Ciro et al., 2025; Ngulube and Molaudzi, 2025; Martínez-Camacho et al., 2025). Professionals are still trying to anticipate the effects (Hamad and Shehata, 2025). We have some suggestions of what the barriers are, such as information quality, privacy, lack of strategy, staff resistance and technical limitations. There seems to be the persistence of a gap between the Global North and the Global South (Buitrago-Ciro et al., 2025). But the potential is there, not only for libraries, but also for access to and preservation of indigenous knowledge (Tella et al., 2025). We are reminded that libraries’ own development happens in wider contexts of policy development (Wilson, 2025).
AI tools, especially commercial AI chatbots, despite their huge potential, have a lot of problems, particularly in terms of accuracy of information (Wang et al., 2025; Elsawy et al., 2025; Rao et al., 2025). Especially with complex professional tasks, they offer some help, but by themselves they are no neat time-saving solution. Using them also shifts librarian roles in potentially significant ways (Rao et al., 2025). In other contexts, we find libraries having made a long-term commitment to developing AI-based services responding to proven need, often based on open-source technologies (Cifuentes-Silva et al., 2025; Jhan et al., 2025). Generative AI has had an impact even here, by lowering the entry-level skills needed to undertake technical development.
A central concern is to educate users in AI literacy, and an intense focus of library activity has been on producing LibGuides about AI chatbot use, reflecting the profession's identity as educators. Our three studies of students show a lot of use in education, but also reveal big gaps in AI literacy (Subaveerapandiyan et al., 2025; Bui et al., 2025; Hossain et al., 2025). Yet it is far from easy to define AI literacy (Ali and Richardson, 2025; Mercer et. al., 2025).
Much of what we read about AI is from the Global North, so it is wonderful to hear the bigger picture of what is happening in different regions. Both the data analysed and authorship in this special issue are truly international. It is heartening to observe several international collaborations underpinning these articles – for example, between Indian and African scholars, between Europeans and authors in South Asia and Latin America, and between Pakistan and Australia (Subaveerapandyian et al., 2025; Hossain et al., 2025; Cifuentes-Silva et al., 2025; Ali and Richardson, 2025). This international scope reflects the strengths of the IFLA community.
The diversity of research methods in use is also impressive. The methods range from literature reviews (including systematic literature reviews), surveys and website analyses to interviews and case studies. Some of our articles feel particularly valuable for the resources they assemble (Tella et al., 2025) or the general applicability of the checklists they offer (Nogueira et al., 2025). The data presented here from surveys and website analyses provides reference points for future longitudinal and comparative studies.
An emerging research agenda
Inquiry into AI in the LIS context will doubtless continue to be an important area of research for some time to come, including the impacts of the next rounds of AI development, such as AI agents. Only over time will we be able to discern patterns in the change of user information behaviour, with its implications for library use. The studies here are all of students – an important group because they are early adopters. Researcher behaviour is also changing. But we will probably soon see widespread changes in everyday user behaviour, with implications for libraries. AI chatbots are changing search expectations. Library service application of AI is in its infancy. Over time, we will get a clearer picture of adoption and variation in use by country and by sector, including some sectors not reflected in this special issue, such as public libraries and corporate and law libraries. This will reveal the evolution of library AI capabilities. It will also set developments in the context of deepening understanding of the environmental and other social impacts of AI.
From this wonderful special issue emerge key questions that will be a long-term agenda for LIS research about AI:
How is user information behaviour and experience changing with AI, differentiating specific user groups such as academic researchers, students and the public? What are the impacts of such changes in use and search expectations on library use? How is AI literacy to be defined, and how does it relate to wider information literacy and digital literacies? How AI-literate are users? How should AI training programmes be designed? How is AI technology and policy evolving, and what are the implications for libraries as entities aligned to wider institutions? What are the ethical issues of AI, including privacy and security, transparency, intellectual property, environmental impact and sustainability, and workforce displacement? How does the evolution and effects of AI technology interact with other long-standing and wider agendas of importance to the LIS community, such as sustainability, decolonisation, and workplace equality, diversity and inclusion? How can AI be applied or be developed for library services, such as in creating metadata generation for discovery or building current-awareness services? What are the drivers and barriers to adoption? What is library AI capability? How is this evolving differently by sector and country? What checklists of criteria are relevant to assessing AI tools? How might AI redefine the library services model, lending libraries a major role in serving communities, enhancing education and transforming research? What are the implications of AI for librarians’ work, knowledge, skills and other competencies, and for professional identities? What are the implications for the library workforce – for example, in terms of demographics and equity? Is there potential for librarians to develop a new value proposition for the future? What are the implications of AI for the training requirements of new librarians and continuing professional development for current professionals? How can AI be used in LIS education?
The IFLA Journal will no doubt address these questions over the next years. This research is likely to be interdisciplinary, exploring AI's impact on information in wider theoretical contexts: with educational scholars around AI literacy; with scholars of business around policy and employability effects; with computer scientists about evolving technologies; with philosophers and social theorists around the wider environmental and societal impacts of AI; with humanists, transforming the nature of their disciplines; and with science, technology, engineering and mathematics scientists, who often pioneer the advanced application of AI, from curing cancer to climate action and autonomous driving. LIS has a lot to offer these wider fields of scholarship gathering around AI.
