Why research impact needs libraries more than ever
Some 58% of researchers are now using AI tools in their work – yet only 22% believe they are currently trustworthy
From Elsevier
Researchers and libraries share a common challenge; both are under increasing pressure to show how research contributes to real-world outcomes. AI is already reshaping the research landscape, transforming how researchers discover literature, evaluate evidence and generate insights.
Research impact measures the way that research advances knowledge, informs practice, supports policy, guides innovation and contributes to wider society benefit. But impact depends on more than speed, it is about trust, transparency and the ability to evaluate evidence well.
The Elsevier Researcher of the Future: A Confidence in Research 2025 report found that 58% of researchers are now using AI tools in their work, up from 37% in 2024. Yet only 22% believe AI tools are currently trustworthy.
Researchers are already using AI but need to determine how to do so effectively and responsibly. Libraries help determine credible tools and showcase relevant use cases to build confidence and enhance researcher impact.
Supporting new research habits
Having access to a tool does not mean knowing how to use it well. As AI is becoming part of the research workflow, libraries have an opportunity to make their strategic value even more visible. The Researcher of the Future report discovered that only 27% of researchers believe they have adequate training in using AI and just 32% say their institution provides good AI governance.
Without guidance, researchers may use tools without understanding their full capabilities, apply them inconsistently or lack the critical evaluation skills needed to assess the quality of outputs. The risk is not only inefficiency but weaker evidence, inconsistent practice and research outcomes that are less credible.
Some researchers may use AI tools to summarise literature without checking the underlying sources. While others may not realise that some AI tools can help compare evidence, identify gaps, find collaborators, discover funding opportunities or produce structured reports for review. Knowledge and adoption will be inconsistent across departments, disciplines and career stages, with researchers at different levels of confidence and familiarity with AI tool capabilities.
Libraries support researchers by focusing on the tasks they need to undertake a research project from scoping, exploring literature, comparing sources, checking claims, identifying collaborators, finding funding and preparing evidence for review. Running workshops and engaging directly with departments, focused on real use cases rather than generic tool demonstrations, builds confidence and competence in equal measure. Libraries are so much more than the resources they provide. Additional value is in how they help researchers turn tools and evidence into trusted and impactful outcomes.
Keeping the human in the loop
AI tools can quickly process large volumes of data, but research impact depends on more than fast outputs. Researchers still need to understand where information comes from, whether evidence is relevant, how claims are supported and what limitations remain.
As an essential human layer in responsible AI adoption, library teams guide researchers in interrogating outputs rather than accepting them at face value, supporting source traceability, evidence evaluation and responsible use. This role closely aligns with what researchers say would increase their confidence: transparent citations, up-to-date scholarly literature and regular human validation of outputs.
Curating tools to strengthen confidence
Libraries can support research impact by helping institutions choose and embed tools that support an integrated research workflow. Institutions have been overwhelmed with how best to manage responsible AI tool adoption with appropriate due diligence and ethical considerations. Researchers are surrounded by AI options, but the challenge is knowing which tools are appropriate, which are secure and which provide the trusted quality needed for credible research outcomes. This has often led to a blanket “no new AI tools’ approach to counteract the mass adoption.
Libraries evaluating tools against those needs can start with asking four questions to help reduce the list to be assessed for a more responsible AI tool list:
- Does the AI tool use trusted scholarly content?
- Does the AI tool provide transparent citations?
- Does the AI tool follow data protection policies?
- Does the AI tool support human oversight?
By bringing more of the research journey into one place, libraries can gather mission-critical insights helping researchers move between discovery, comparison, synthesis, funding exploration and collaborator discovery with greater clarity.
Tools like LeapSpace combine responsible AI with peer-reviewed, publisher-neutral full text and abstracts, trusted data sources and human oversight. The shift to integrating responsible AI into the core research process sees an AI tool move from a separate assistant to a core part of the research workflow. For libraries, the significance is not only the tool but what it represents: curation, implementation and guidance becoming a central pathway to research impact.
Supporting evidence impact
The research lifecycle is embedded across the institution and stakeholder communities. Libraries are the connecting point with a cross-institutional role. Their strategic value in AI tools adoption supports that idea that AI tool adoption is not only a technological decision, but also a research culture decision. Libraries are part of the story to help determine and communicate impact responsibly, and ethically. Libraries already make impact easier to demonstrate through research visibility, responsible metrics, literature searching and researcher profiles. As AI tools become more embedded in the research workflow, libraries can ensure impact narratives remain credible, transparent and grounded in trusted quality.
The way research impact is measured is also evolving. Bibliometrics alone no longer tell the full story, and AI is accelerating that shift. Libraries that stay ahead of emerging frameworks for assessing research quality and influence can offer guidance that goes beyond tool use, helping institutions demonstrate impact in ways that are credible, transparent and aligned with where the field is heading.
Research impact needs libraries
AI-enabled research tools can help researchers save time, but also explore evidence, identify connections and move from question to insight more efficiently. But the value of those tools depends on how fully, responsibly and confidently they are used.
Libraries share knowledge, improve impact and build confidence by guiding researchers in what AI can do, how to use it critically and with integrity to support stronger research outcomes. AI may help research move faster but library teams help ensure it moves with trust, transparency and purpose, advancing human progress together.