Can researchers achieve real-world impact with AI-assisted research?
AI tools can allow researchers to move faster through discovery, comparison, synthesis and exploration
From Elsevier
Expectations around research impact are changing, not only around what research achieves, but how that impact is created, evidenced and trusted. The Researcher of the Future: A Confidence in Research 2025 report, from 3,234 active researchers and leaders across 113 countries, found that:
Some 67% agree there is more emphasis on mission-oriented research
Research impact is more than journal metrics, it is about how research advances knowledge, informs practice, supports policy or contributes to wider benefit. But getting there can be slow. Researchers are working with higher volumes of information, increasing pressures to publish, uncertainty around funding and rising expectations to show the relevance of their work.
The Researcher of the Future report states:
Just under half of researchers (45%) agree they have sufficient time for research, while two thirds (68%) say the pressure to publish is greater
Couple that with a research workflow that can be siloed and fragmented, literature discovery, funding searches, finding collaborators, comparing articles, validating claims, reviewing source material, creating summaries and refining ideas means multiple channels, tools, systems and files. This causes inconvenience, frustration, and lost time when researchers need to be evaluating evidence, thinking critically and deciding what comes next.
AI adoption is an opportunity
AI tools can allow researchers to move faster through discovery, comparison, synthesis and exploration when adopted with appropriate governance, user responsibly, and aligned to research objectives. Using AI to support researchers in time-consuming activities means their time is better spent researching, validating their research question and understanding the integrity of the work used to inform their output. The most common current uses are directly connected to the research workflow:
- 61% use AI to find and summarise the latest research
- 51% use it to perform literature reviews
- 41% use it to draft grant proposals
- 38% use it to analyse research data
The opportunity is not just to make individual tasks faster but to bring more of the research journey into one place. Researchers can explore a complex question, compare evidence in articles and perspectives, check the evidence behind claims, surface patterns and gaps across the literature, identify potential collaborators, discover relevant funding opportunities and produce a structured report for review. This means spending less time moving between disconnected systems and more time interpreting what the evidence means.
It’s worth noting, confidence in AI is still developing. Only 22% of respondents in the Researcher of the Future report believe AI tools are currently trustworthy. However, researchers say their confidence would increase through trust markers including transparent citations, up-to-date scholarly literature, training on high-quality peer-reviewed content and regular human validation of AI outputs. In other words, AI should not ask researchers to trust a black box but help researchers inspect the evidence more efficiently.
Purpose-built research tools, like LeapSpace, are designed to support researchers in working faster, smarter and thinking deeper by combining responsible AI and trust indicators with peer-reviewed, publisher-neutral full text and abstracts, and include trusted data sources, transparent processes and human oversight. This shift integrates an AI tool into the research workflow rather than as a separate assistant.
But AI tools, however well designed, are not replacements for people. They do not remove the need for independent evaluation. They do not make impact automatic. They support researchers by helping them complete important tasks more effectively, while keeping responsibility for interpretation, judgement and direction where it belongs, with the researcher.
Governance is the foundation
To support the research community, institutions are tackling how to best manage responsible AI tool adoption and undertake appropriate due diligence and ethical considerations. AI tools foregrounding responsible AI guardrails alongside powerful functionality can help to assess quality, transparency and reliability of outputs.
Bringing the research journey into one place, from trusted sources with responsible tools helps support the vital step of human judgement, adding another governance layer. This governance can help researchers focus on more defensible and impactful results.
Institutions and libraries have a critical role to play
Confidence and adoption can be low in AI when only 27% of researchers believe they have adequate training in using AI and just 32% say their institution provides good AI governance according to the Researcher of the Future report. Researchers need guidance to understand which tools to use, support responsible use, build AI literacy and evaluate evidence from outputs.
Libraries already play a significant part in supporting researchers; they are essential. Incorporating AI use and governance in knowledge sharing with ethical considerations is a natural fit for libraries continuing to demonstrate relevance and impact in their own right. That will matter more as AI embeds itself into everyday research practice, helping researchers and institutions demonstrate impact in a more measured and analytical way.
The real promise of AI for research impact is not speed for its own sake but turning authoritative research into outcomes that matter. If AI-enabled tools are going to support better research outcomes, researchers need more than access.