OU secures major grants to strengthen trust in AI

A researcher from The Open University's Knowledge Media Institute (KMi) has secured a major UK Research and Innovation (UKRI) grant to help ensure artificial intelligence can be used safely, fairly and transparently in research assessment.

Dr Francesco Osborne will lead TRUST-AI-PR (Trustworthy and Robust AI Support for Peer Review), an international project worth approximately £300,000, including around £250,000 for The Open University. The project will develop the next generation of trustworthy AI systems for peer review, helping research organisations make informed decisions about the use of AI in evaluating research proposals and publications.

As the use of generative AI continues to expand across the research sector, publishers, conference organisers and funders are increasingly exploring how AI could support research assessment. While these technologies have the potential to improve efficiency and reduce workloads, concerns remain about whether AI-generated assessments can be manipulated or biased.

TRUST-AI-PR will address these challenges by developing technical safeguards against manipulation and translating the findings into practical guidance for organisations using AI-assisted peer review. The project will investigate how hidden instructions, persuasive language and targeted edits can influence AI-generated assessments, while also testing methods to improve robustness and transparency.

Working with partners including DORA, Springer Nature, Jisc, the International Association of Scientific, Technical and Medical Publishers (STM), and Fundación "la Caixa", the team will develop an assurance framework and implementation toolkit to help organisations deploy AI responsibly in research assessment.

Driving trustworthy innovation in research

Dr Osborne, said:

"AI is set to become central to how we conduct and evaluate research. The choices we make now will shape how ideas are judged and which research receives support. We must build fairness, robustness and transparency into both the technologies and the practices that govern their use, while preserving human judgement.

"TRUST-AI-PR will turn these principles into practical tools and guidance, grounded in rigorous scientific testing and the expertise of partners across the research sector.”

Funded through UKRI's Metascience Research Grants Round Two, the project will begin in December 2026 and run for 18 months in collaboration with the University of Cagliari. KMi's Angelo Salatino will oversee benchmark development, while international co-lead Diego Reforgiato Recupero will lead work on developing safeguards against manipulation.

The award forms part of a wider programme of research led by Dr Osborne's team exploring how AI can improve research evaluation and funding. Alongside TRUST-AI-PR, the team has secured two additional UKRI-funded projects, worth approximately £200,000, focused on AI-supported grant assessment and understanding the long-term impact of research funding. Together, these projects reinforce The Open University's leadership in shaping the future role of AI in research and innovation.