Building Trust in AI-Era Research Assessment
Key Takeaways from the CoARA1–EC–ENRIO-EDI-TUM Conference (Brussels, 2 June 2025)
On 2 June 2025, stakeholders from across the globe gathered in Brussels and online for the high-level conference "Advancing Responsible Research Assessment for Funders in the European Digital Space. A dialogue on ethics, integrity, and open infrastructure." The event addressed the urgent need to reform research assessment systems in light of rapid digital transformation, artificial intelligence, and shifting global expectations for openness and fairness.
The conference was co-organized by multiple actors: the CoARA Working Groups on "Towards Open Infrastructures for Responsible Research Assessment (OI4RRA)" and "Ethics and Research Integrity Policy (ERIP)", the Ethical Data Initiative, European Network of Research Integrity Offices (ENRIO), the European Commission and TUM Think Tank. OpenAIRE, as co-chair of the OI4RRA WG, played a key role in shaping and leading several sessions.
The day opened with a keynote by Mihalis Kritikos, Secretary of the European Group on Ethics in Science and New Technologies (EGE) at the European Commission (DG RTD). His talk, "Ethics in Action: From Principles to Practice – Operationalizing Ethics for AI-Ready Research Assessment" offered a bold vision for embedding ethics as a guiding force in all AI-enabled decision-making. Kritikos argued that ethics is not merely a safeguard, it is a political deliverable and must be built into the infrastructures that shape how knowledge is created and evaluated.
Spotlight on Session 1: Ethics and Trust in Research Assessment
Moderated by Angeliki Tzouganatou (OpenAIRE), Session 1 explored how ethical principles can and must be embedded into modern research assessment systems, especially as digital tools and AI increasingly influence evaluation processes.
Perihan Elif Ekmekçi (TOBB ETÜ, CoARA-ERIP) made the case that collaboration, inclusion, and ethics should form the core of assessment design, not act as retroactive justifications. Giovanna Lima (DORA) emphasized the shift towards narrative-based assessment, transparency, and trust, while Bianca Kramer (Barcelona Declaration) urged a stronger link between open infrastructure and epistemic justice. Bert Seghers (ENRIO) and Susanne van den Hooff (University of Humanistic Studies) underlined the need for infrastructures that are by default ethical, not reactive. The use of opaque algorithms or inaccessible metrics, they warned, undermines both transparency and fairness.
Mihalis Kritikos returned to the debate with a strong message on AI literacy and democratic oversight. He argued that "trust" is not just a social value, it is a system output that depends on how technologies are designed, governed, and applied.
The session collectively demonstrated that ethical, human-centered design is essential to making AI-enabled research assessment credible, inclusive, and trustworthy.
Spotlight on Session 2: Responsible by Design – Open Infrastructure
Natalia Manola (CEO of OpenAIRE, Co-chair of the CoARA OI4RRA WG) opened Session 2 with a keynote that challenged the research community to confront a critical paradox: while science increasingly relies on open infrastructures to produce knowledge, its assessment still depends on closed, opaque systems that restrict access, entrench silos, and obscure accountability.
She argued that open infrastructures are not optional, they are essential to ensuring trust, equity, and transparency in research assessment. From software and datasets to mentoring and community impact, open infrastructures enable institutions and funders to capture and evaluate the full spectrum of research outputs in ways that reflect societal value.
Manola introduced the OI4RRA WG's Framework and Conceptual Architecture, a four-tiered, interoperable model for transparent, machine-actionable evaluation, along with practical action plans for research-performing (RPOs) and research-funding organizations (RFOs). These tools, now available on Zenodo, offer clear pathways to operationalize responsible research assessment.
Her central message was clear: we need full-stack investment in open infrastructures, from PIDs and APIs to governance and sustainability. These systems must be treated as public knowledge assets, not short-term projects. Manola called for new funding models grounded in shared ownership, distributed responsibility, and a long-term commitment to openness; responsible by design.
Looking Ahead
The momentum will continue at the High-Level Conference on Reforming Research Assessment, to be held in Copenhagen on 3–4 December 2025 under Denmark's Presidency of the Council of the EU, with Aalborg University and international partners. David Budtz Pedersen, Chair of the Program Committee, invites the community to translate the ideas launched in Brussels into actionable frameworks for reform.
Outlook: Ethics, Equity, and Collective Impact
The closing dialogue of the conference made one thing unmistakably clear: reforming research assessment is not just a technical upgrade. It is a collective commitment to reposition science as a public good, grounded in ethics, inclusion, and transparency.
Participants reflected on the shared language that must underpin any future system, ethics and responsibility. These are not compliance checkboxes but the glue that holds together trust, collaboration, and legitimacy. As one speaker put it, communication isn't just part of the challenge, it is the challenge, and perhaps our most powerful tool.
But communicating ambition is not enough. There was a sobering recognition that local institutions, where much of the reform must happen, are often unclear on how to act. The call was for guidance that is not only strategic but also clear, actionable, and digestible. Bridging global aspirations with local implementation will require a shift from jargon to shared understanding.
The future of research assessment will also depend on AI literacy and investments in open, publicly governed infrastructures. Only by developing the capacity to understand, shape, and trust these systems can we ensure that digital tools serve the values we claim to uphold.
Finally, the conference ended with a powerful assertion: ethics is not neutral, and infrastructures are not passive. Both are designed, and they reflect who we are as a community. Reforming assessment, then, it is not just about improving performance indicators. It is about redefining what—and who—counts in science, and aligning our tools, incentives, and cultures accordingly.
1. This conference is organized on behalf of CoARA by its working group on Open Infrastructures for Responsible Research Assessment (OI4RRA) and its working group on Ethics and Research Integrity Policy for Responsible Research Assessment in Data and Artificial Intelligence (ERIP). The conference receives funding support from EU funding number (Grant Agreement ID) for the CoARA Boost project is 101131826, as well as individual travel support from the Directorate-General for Research and Innovation (DG-RTD), European Commission.
When you subscribe to the blog, we will send you an e-mail when there are new updates on the site so you wouldn't miss them.
