Liviu Știrbăț | Head of AI in Science at the DG for Research and Innovation in the European Commission

How Europe is moving from an AI in Science strategy towards an enabling framework for AI-powered research in practice. The keynote will provide the broader policy context for the discussions that follow, exploring how Europe is seeking to accelerate the responsible uptake of AI across the scientific process by bringing together talent, data, computing capacity and research funding, while at the same time safeguarding scientific integrity, methodological rigour and human responsibility. Particular attention can be given to the transition from policy ambition to implementation through RAISE, the ERA Action on AI in Science and the evolving Living Guidelines on the responsible use of generative AI in research.

Against this broader landscape, the keynote will provide a policy framework for the conference’s two subsequent roundtables on how AI is beginning to change research practice itself – including the way researchers develop ideas, produce scientific content, interact with funding systems and participate in peer review and evaluation.

Barbara Świątkowska, National Science Centre, Poland

AI is beginning to unsettle long-familiar habits in proposal writing — not by replacing researchers, but by reshaping how ideas emerge, take form and become fundable narratives. As tools accelerate drafting, summarising and structuring, questions of ownership, originality and intellectual contribution move to the foreground, echoing the recent EU Guidelines on the Responsible Use of Generative AI in Research, which emphasise transparency, human oversight and clear documentation of how AI is used in scientific work. Institutions now face a new balancing act: supporting researchers who navigate rising workloads and proposal pressure while also safeguarding integrity, accountability and responsible practice. This panel asks how the research system should respond to these shifts, and what questions should bring into Europe’s broader discussion on AI in proposal development.

■ Michal Kompan, Research Director at the KInIT, Slovakia

■ Michaela Liegertová, Vice-Dean at the UJEP, Czechia

■ Sophie Pornschlegel , European Policy Centre, University of Maastricht

Moderator: Barbara Świątkowska, National Science Centre, Poland

Rather than revisiting technical aspects, the conversation centres on shared experiences — how AI is already appearing in proposal development and how these changes are perceived across different parts of the system. The goal is to surface the perspectives, questions and concerns emerging in the V4 region, whether they relate to originality, fairness, workload, skills or responsible practice. By bringing together researchers, EU institutions and NCPs, the session encourages participants to articulate what feels promising, what feels risky and what remains uncertain within their national and institutional contexts.

AI is also beginning to influence how proposals are read, compared and judged — not by replacing evaluators, but by subtly reshaping the assessment process itself. Tools that summarise, cluster or highlight content can speed up review, yet they also raise questions about fairness, consistency and the evaluator’s own judgement. The recent EU guidelines on responsible AI use emphasise human oversight, transparency and accountability, reminding us that evaluation must remain a human-led process even when supported by automated systems. As proposal volume grows and expectations rise, the research system must consider where AI genuinely helps and where it risks introducing new biases or dependencies. This panel explores these tensions.

■ Margaret Ohia-Nowak, Proffesor at UMSC Lublin

■ Martin Weis, Professor at the SUT in Bratislava, Slovakia

■ Miroslav Havránek, Member of Presidium of the Technology Agency of the Czech Republic

Moderator: Miguel Ruiz Oses, DG Research and Innovation, European Commission

AI is beginning to reshape how proposals are read, compared and judged, offering tools that can support evaluators under growing pressure - yet also raising concerns about fairness, consistency and the limits of human judgement. The roundtable examines where AI provides meaningful support in evaluation workflows and where it risks exerting undue influence. Participants will reflect on how evaluators interpret AI-generated summaries or highlights, how national systems perceive the boundaries of responsible use, and what pressures arise as proposal volumes increase. By bringing together V4 researchers, NCPs and institutional actors, the session creates space to articulate concerns and priorities.