From Access to AI-Readiness: The Evolving Role of Research Libraries in Open Science

Research libraries matter more than ever. As trust in knowledge comes under pressure, research systems grow more complex, and AI reshapes how information is created, discovered and reused, libraries have a defining role to play.

This is a sponsored post from LIBER Gold Sponsor Frontiers, written by Liz Bowley.

Reflecting the LIBER 2026 theme, The Power of Libraries in an Uncertain World, Liz Bowley, Director of Publishing Development at Frontiers, explores how research libraries are evolving from access support to knowledge stewardship, and why their expertise will be essential in building open science systems that are trusted, reusable and ready for AI.

Research libraries have always helped knowledge move further. They make research easier to find, access, preserve and reuse, and help researchers navigate publishing, copyright, open access and data sharing.

Today, that role is expanding. As research becomes more data-intensive, interconnected and shaped by artificial intelligence, libraries are helping institutions build the skills, standards and infrastructure needed to make research data more structured, transparent, reusable and trusted.

Open science is moving beyond access

Open science has transformed research by making knowledge more widely available. But open access is only part of the story. The next challenge is to ensure research is not only available, but usable.

Since their introduction in 2016, the FAIR principles have helped shape expectations for how research data should be managed, shared and reused. They call for digital research assets to be findable, accessible, interoperable and reusable, with the data, methods and context needed to make research meaningful.

In the age of AI, this matters even more. Research data must not only be FAIR. It must also be AI-ready and responsibly governed.

AI can only support discovery if the knowledge systems around it are standardized, reliable and robust. That makes research libraries central to the next phase of open science. They sit at the intersection of research practice, information literacy, scholarly communication and institutional policy, and can help ensure open research is discovered, interpreted and reused responsibly by people and machines.

AI-readiness depends on library strengths

AI-readiness is often discussed as a technical challenge. But at its core, it depends on principles that libraries already understand deeply.

AI-ready research needs high-quality metadata, clear provenance and reliable links between publications, datasets, methods, funding, authorship and institutional context. It needs preservation, discoverability, rights clarity and responsible reuse.

These are core parts of library expertise. What is new is the scale and urgency.

AI systems depend on the quality of the information they process. If research outputs are poorly described, disconnected or inconsistent, AI systems will struggle to reuse them responsibly. If data lacks context, provenance or structure, the risks increase. If rights and licensing are unclear, trust and compliance become harder to maintain.

From access support to knowledge stewardship

The role of the research library is no longer only to provide access to research outputs. It is to help ensure knowledge can be trusted, connected and used responsibly.

That includes supporting researchers with data management plans, repository deposits, metadata quality, open access publishing, rights and licensing, persistent identifiers, FAIR principles, responsible reuse and, increasingly, AI literacy.

The bigger question is how to make responsible data sharing easier for researchers, not harder. Too often, data is not shared effectively because the administrative burden is too high. Researchers must navigate policies, metadata requirements, repository choices, licensing questions and compliance expectations without enough practical support. At the same time, institutions need clear ways to demonstrate compliance with funder and policy requirements, from usage reporting and mandate tracking to transparent, audit-ready records.

Libraries are well placed to reduce that burden because they understand both research practice and the information systems that make knowledge discoverable, reusable and compliant. They can help align AI policies with open science principles, support stewardship and accountability, and work with partners to build the reporting and transparency institutions need to show that open science is being put into practice.

FAIR² and the move towards AI-ready open science

FAIR² Data Management builds on the FAIR principles and extends them for an AI-enabled research environment.

Where FAIR focuses on making data findable, accessible, interoperable and reusable, AI-ready open science also requires data to be readable by people and machines, with its provenance preserved, its quality checked and its context intact.

In practice, that means publishing data alongside a peer-reviewed data article, so the methods are described, reviewed and linked to the dataset. It means recording provenance in a machine-readable form, connecting data through persistent identifiers to authors, funding and related research, and placing responsible governance, clear rights, disclosed AI use and human oversight at the heart of reuse.

This matters because open science cannot reach its full potential if research data remains fragmented, inconsistent or difficult to reuse. For libraries, it speaks to a growing institutional need: helping researchers meet expectations for open, reusable and well-managed research outputs while preparing for an AI-enabled future.

It also underlines the importance of collaboration. Publishers, libraries, institutions, funders, researchers, infrastructure providers and technology partners all need to work together.

Libraries can help protect trust in the academic record

As AI use grows, questions of trust will become more urgent. Can a research output be traced back to its source? Is the dataset reusable? Are the methods clear? Has AI use been disclosed where appropriate? Are citations real and verified? Is metadata complete? Can the work be discovered, interpreted and reused responsibly?

These questions sit at the heart of research integrity. They are also library questions. When methods are peer-reviewed and linked to the data, when provenance is recorded in a machine-readable form, and when persistent identifiers connect a dataset to its authors, funders and context, the research record becomes easier to trace, verify, interpret and reuse.

A defining moment for research libraries

As AI use grows, the importance of research libraries will not diminish. It will increase. As research becomes more data-intensive, automated and interconnected, trusted stewardship will become even more essential.

Libraries can help ensure AI supports openness rather than opacity, and that automation strengthens research quality rather than undermines it. By working with publishers, institutions and infrastructure partners, they can help shape open science systems that are rigorous, responsible and easy for researchers to adopt.

Open science has always depended on more than access. It depends on trust, context, infrastructure and human expertise. In the age of AI, those foundations matter more than ever.

To find out more about how your institution can support sustainable open access to science, please visit: https://www.frontiersin.org/about/open-access-agreements