AI and Librarians: What's Actually Happening and What to Do

The honest assessment

Librarianship has a curious relationship with AI. On one hand, the core historical function of a librarian — helping people find information — is precisely what AI claims to do. On the other hand, modern librarianship has evolved well beyond information retrieval into community building, digital literacy education, and knowledge curation that AI is nowhere near replicating.

Let's deal with the information retrieval piece first, because it's the most obvious challenge. When someone needs to find information on a topic, the first instinct in 2026 is not to visit a library or ask a librarian. It's to ask ChatGPT, search Google, or use Perplexity. AI chatbots can answer reference questions instantly, summarise research topics, recommend reading lists, and even guide someone through a research methodology. For the casual information seeker, AI is faster, more convenient, and available 24/7. This is real, and pretending otherwise doesn't help.

AI is also making inroads into the technical library functions. Cataloguing and metadata creation — historically a significant part of library work — can be substantially automated. OCLC and Ex Libris have been developing AI-powered cataloguing tools. AI can classify materials, generate subject headings, and create metadata records from analysing content. Collection development algorithms can analyse usage data and recommend acquisitions. Discovery layers in library management systems increasingly use AI to improve search relevance and recommend related resources.

But here's where the picture gets more interesting. Libraries have spent the last two decades evolving from "places where you borrow books" into community hubs that provide digital access, literacy programmes, community meeting space, children's educational activities, job-seeking support, mental health signposting, and safe spaces for vulnerable people. A public library in 2026 is as much a social service as an information service. AI doesn't provide a warm, safe space for a homeless person on a cold day. AI doesn't run a reading group for isolated elderly people. AI doesn't teach a retired steelworker how to use a computer to apply for benefits.

Academic and research librarianship has also evolved. Research librarians help academics navigate complex databases, develop search strategies for systematic reviews, manage research data, advise on open access publishing, and support digital scholarship. These require a combination of domain knowledge, research methodology expertise, and interpersonal skills that AI assists with but doesn't replace.

Your exposure level: Medium

Medium. The information retrieval and cataloguing aspects of the role face significant AI exposure. The community, educational, and curation aspects face very little.

The medium rating reflects this split. If your role is primarily answering reference questions and cataloguing materials, your exposure is higher than medium. These are the functions most directly replicated by AI. If your role involves community programming, digital literacy education, research support, collection curation with a point of view, or managing a physical library space, your exposure is lower.

The structural context matters too. Public libraries in the UK have faced over a decade of funding cuts that have already reduced the workforce significantly. CIPFA figures show that the number of paid library staff fell substantially between 2010 and 2024. AI isn't the primary threat to library jobs — austerity is. But AI could provide further justification for reducing staff if library services can be partially automated. Conversely, AI could free up remaining staff to focus on the high-value community work that justifies public library funding.

Academic libraries face a different dynamic. University budgets are under pressure, and library services are often seen as overhead rather than core academic function. AI tools that allow students and researchers to find information independently could reduce demand for reference services. But the growing complexity of research data management, digital scholarship, and information literacy education creates new demand for librarian expertise. The role is changing. It's not disappearing.

The 90-day action plan

1. This week: become fluent in AI information tools. Use ChatGPT, Claude, and Perplexity extensively. Understand what they do well and what they get wrong. Try asking them the kinds of reference questions your users bring you. Note where they give accurate, well-sourced answers and where they hallucinate, give outdated information, or miss crucial nuances. This knowledge is essential because your users are using these tools, and you need to be the person who can help them use them well.

2. Week two: develop AI literacy teaching materials. Create a simple guide or workshop on how to evaluate AI-generated information. Cover hallucinations, source verification, bias, and the kinds of questions where AI is reliable versus unreliable. This is one of the most valuable things a librarian can do right now. The world is drowning in AI-generated content, and people desperately need help learning how to navigate it critically.

3. By day 30: propose an AI literacy programme. Whether you're in a public library, academic library, or special library, propose a structured programme teaching AI literacy to your users. For public libraries, this could be "using AI safely and effectively" sessions for different age groups. For academic libraries, this could be guidance on appropriate AI use in research. You're positioning yourself as the expert in navigating the AI information landscape.

4. By day 45: explore AI tools for collection management. Look into how AI can help with collection analysis, acquisition recommendations, and usage pattern insights. If your library management system has AI features, learn them thoroughly. The librarian who can use AI to make better collection decisions and demonstrate the value of the library through data is in a strong position.

5. By day 60: strengthen your community programming. Develop new programmes that leverage what libraries uniquely offer: physical space, community connection, and trusted expertise. Digital creation workshops. Author events. Reading groups. Coding clubs. Community information sessions. These are the activities that justify libraries as community institutions and are impossible to automate.

6. By day 75: develop digital scholarship or data skills. Depending on your setting, learn about research data management, digital preservation, data visualisation, or digital humanities tools. These are growing areas where library expertise is valued and where AI creates more work (more digital content to manage) rather than less.

7. By day 90: articulate the modern library case. Whether to your management, your local authority, or your university, make the case for the evolved library. "Our library provides AI literacy education, community programming, digital access for the digitally excluded, research support, and curated information services. AI makes some of this more efficient, but it also makes the educational and community elements more important than ever."

AI tools you should be using this week

• ChatGPT for Work — Use it the same way your users do, so you understand its capabilities and limitations from the inside. Also useful for drafting programme descriptions, creating promotional materials, writing reports, and generating ideas for events and activities. Ask it to help you write a grant application or a business case for a new library service.

• Claude for Work — Strong for research-oriented work. Use it to help develop systematic review search strategies, analyse research questions, and create literature summaries. Claude's tendency to be careful and acknowledge limitations makes it particularly suitable for academic and research library applications where accuracy matters more than speed.

• Perplexity for Research — This is the AI tool most directly relevant to reference work. It searches current sources and provides referenced answers. Use it alongside traditional reference tools to compare results. Understanding where Perplexity excels and where it falls short makes you better at helping users navigate AI-assisted research.

• Microsoft Copilot for Work — Helps with the administrative burden of library management. Reports, emails, meeting summaries, programme documentation, and data analysis in Excel. Public librarians in particular spend significant time on administration and reporting that Copilot can accelerate, freeing up time for user-facing work.

What to say in meetings

When leadership questions the need for librarians when AI exists: "AI can answer factual questions, but it can't verify its own accuracy, teach someone how to evaluate information critically, curate a collection with a community's specific needs in mind, or provide the safe physical space that our users rely on. What we're seeing is that AI makes our educational role more important, not less. People need help navigating a world with more information and less certainty about what's reliable."

If colleagues are worried about AI replacing reference services: "AI handles the simple lookups well. It's the complex, nuanced, and critical research support where we add the most value. The question isn't whether people can find information without us — they always could. It's whether they can evaluate, contextualise, and apply that information effectively. That's our expertise, and AI actually makes it more relevant."

In budget discussions: "Our library now provides AI literacy education, digital inclusion support, community programming, and research services alongside traditional collections. Usage of our [workshops/programmes/spaces] has [grown/remained strong]. I'd like to develop our AI literacy offering further because it addresses a genuine community need that no other service is meeting."

If the worst happens

If your library role is cut — and given ongoing funding pressures, this is a real possibility regardless of AI — your skills transfer to a surprising range of roles. You can organise and classify information. You can research efficiently. You can teach and communicate. You understand data management. You're comfortable with technology. You can work with diverse communities. These skills transfer to information management, knowledge management, data management, education, community development, and research.

Adjacent roles to consider: information manager in a corporate or legal setting, knowledge manager, research analyst, community development officer, digital inclusion coordinator, data manager, content curator, or education coordinator. Many librarians also move into publishing, archiving, museum work, or policy research. The combination of information expertise and people skills is valued more broadly than you might think.

Here's what i believe about libraries and librarians. The profession has survived every technological disruption it's faced — microfilm, databases, the internet, Google, and e-books. Each time, pessimists predicted the end of libraries, and each time, librarians evolved. AI is the latest disruption, and yes, it's significant. But the fundamental need that libraries serve — helping people access, evaluate, and use information, and providing community space and services — isn't going away. It's more important than ever in a world awash with AI-generated content of variable quality. Be the person who helps people navigate that world. That's not a role AI can fill.