Artificial intelligence (AI) is now a present force, reshaping the landscape of higher education and scholarly communication as it unfolds. For academic libraries, the implications are profound. On one hand, AI holds the potential to accelerate discovery, streamline metadata curation, and empower research analytics. On the other, it brings new questions of ethics, reliability, copyright, data governance, and equity.

The message from the academic community is clear: librarians need trusted partners who can align with their values and help navigate this rapidly evolving landscape.

Navigating AI’s promise and peril

Recent studies show how quickly AI is embedding itself into research and teaching. Ithaka S+R surveys from 2024 and 2025 highlight widespread faculty experimentation with generative AI yet also lingering uncertainty about how to integrate it pedagogically. Meanwhile, Jisc’s 2025 report reveals that students are embracing AI tools but are calling for clear guidance, equitable access, and concern over effects on skills and employability.

But alongside opportunity comes risk. A 2024 audit in Nature Machine Intelligence assessed over 1,800 text datasets and found rampant mislabeling and missing licensing — raising ethical and legal concerns. Similarly, MIT researchers reported in August 2024 that more than 70 percent of datasets used to train language models omit licensing information, while around 50 percent contain errors. Such data opacity undermines essential values of traceability and trust for libraries.

At the same time, policymakers are raising the bar. The EU AI Act, entering into force in stages through 2025-2027, mandates that providers of general-purpose AI disclose summaries of their training data, including sources and licensing, before deployment. Libraries must partner with vendors who embrace transparency and accountability.

What trusted partnerships looks like

We believe AI must serve academia, not the other way around. Libraries will be critical in setting standards, evaluating tools, guiding researchers and students, and ensuring that technology adoption does not compromise equity, privacy, or scholarly rigor.

But they cannot do it alone. They need partners who share their values, who bring transparency to both algorithms and data, and who will stand with them in shaping responsible innovation.

Through discussions with librarians, one theme consistently emerges: not all AI tools are created equal. To support the academic mission, partners must deliver more than technology, they must embody the values of the scholarly community.

Key attributes include:

• transparent models and data: openness about training sources, licensing, and documented biases.

• responsible innovation: tools that augment but do not replace human expertise.

• ethical stewardship: strong safeguards for privacy, equity, accessibility, and research integrity.

• alignment with academic standards: adherence to values articulated by ARL, IFLA, and ALA, and readiness for EU AI Act requirements.

How Frontiers is supporting libraries

As a fully open access publisher, Frontiers is committed to partnering with libraries on this journey. Our approach to AI is grounded in transparency, responsibility, and respect for scholarly values.

• Integrity first. Our in-house AI assistant, (AIRA) supports editors by running more than 40 checks for duplication, manipulation, and plausibility, while ensuring manuscripts remain secure and confidential under human oversight.

• Clear guardrails. We require disclosure of AI use in research submissions and prohibit uploading manuscripts into external AI tools, protecting confidentiality and trust in peer review.

• FAIR-aligned data practices. We embrace the FAIR (findable, accessible, interoperable, reusable) principles, advocating for provenance-rich, AI-ready datasets that reinforce the academic record.

As AI reshapes the world of research, libraries are the guides keeping scholarship grounded, but their impact grows when publishers step in as committed collaborators.

By offering tools, expertise, and transparent practices, publishers like Frontiers help libraries navigate complexity, implement responsible AI, and empower researchers with confidence. In this partnership, both libraries and publishers play a vital role in shaping an AI-enabled future that strengthens research, supports learning, and preserves the trust at the heart of academia.

About Frontiers 

Frontiers is one of the world’s largest and most impactful research publishers, dedicated to making peer-reviewed, quality-certified science openly accessible. With more than 3 million researchers across 222 community-led journals covering 1,700 academic disciplines, we provide researchers with a trusted, cutting-edge, AI-powered open science platform to rigorously review their findings and maximize the dissemination of their discoveries.​ As an open access pioneer, we actively drive the global transition to open science, working with researchers, universities, educators, policymakers, and businesses.

For more details about our open access agreements, please visit our website. If you want to learn more about our institutional partnerships, submit this contact form or email us at institutions@frontiersin.org.