Facilitated cooperation for AI in Science (CSA)

Expected Outcome:

Projects are expected to contribute to the following outcomes:

• Identify the long-term research challenges where AI can make a meaningful breakthrough contributing to EU’s competitive edge in selected scientific disciplines/areas, through a Strategic Research and Innovation Agenda.
• Provide evidence to structure the resources for AI in Science at European level, as a feasibility test towards potential R&I initiatives beyond the CSA that could optimise access to relevant data, infrastructure and talent across different scientific domains for more and better AI-enabled research.
• Coordinate, strengthen the network and raise awareness and a community of scientists, including citizen scientists, research organisations and stakeholders towards new paradigms of research with AI.

Scope:

Artificial intelligence is a game-changer for science and innovation, and promises significant opportunities to boost the European competitive edge in R&I that need to be capitalised on. The aim of the CSA is to structure AI-enabled research in Europe and assess options towards optimising the ecosystem for AI in Science in Europe, through a Strategic Research & Innovation Agenda and assessing the potential for possible future R&I initiative(s), in line with the recommendations of the Scientific Advice Mechanism^[1], and the European Commission President’s political guidelines, for the setting up of an AI Research Council^[2]..

The project should develop a Strategic Research and Innovation Agenda for AI in Science by mobilising large groups of domain and AI researchers in different fields to identify key long-term research challenges in a diverse range of scientific areas where AI can make a meaningful difference for scientific breakthroughs, which are compelling to the EU competitive, environmental and social policy agenda. The project should come up with pilot areas from across Horizon Europe Pillar II Clusters, building on Europe’s competitive advantages in science and AI technologies. The research challenges should be related to prediction and design problems in the different scientific fields identified that could be solved with AI.

The SRIA should also include areas where AI can improve generic scientific tasks e.g virtual research assistant / tools for literature-based discovery, for improving / enabling research workflows, lab automation and collaborative human-AI work in science. Research priority areas involving the use of models based on frugal AI, which are more compact, more efficient and less energy intensive, as well as human-centric and trustworthy AI for scientific work should also be explored.

The project should build evidence and assess the needs and potential for R&I initiatives for AI in Science beyond the CSA, in an effort to identify ways to improve the EU landscape for support for AI in Science, to be discussed and agreed upon with the Commission and the Member States. The assessment should identify the ways for improving data access, infrastructure and support services, as well as skills and talent-related needs to boost the integration of AI in different fields of science at larger scale in Europe in research processes and lab automation, while promoting reproducibility, transparency and open science. It should also identify options for EU to better enable cooperative development and sharing of AI models for scientific discovery across different scientific fields. It should also take into account existing EU efforts to support access to data, research infrastructures, networks, HPC.

Different scenarios of R&I initiatives and infrastructure improvements should be prototyped together with a diverse range of users and stakeholders from the research community, industry, start-ups, civil society and policy-maker communities. Based on the feasibility test results, the project should develop a roadmap on the needed steps for a more effective coordination between the domain and AI scientific communities in Europe and the needed upgrades in service and infrastructure provision at EU level for the integration of AI in different scientific fields, including research processes engaging citizens and civil society, e.g. Citizen Science.

The proposals should also provide coordination and dissemination for interdisciplinary AI-enabled science to facilitate stakeholder engagement, coordination and promotion of AI in Science initiatives across Europe. The CSA should develop a website, organise awareness raising events for the benefits of AI in Science and create opportunities for exchanging on good practices.

Projects should build on or seek collaboration with existing projects and develop synergies with other relevant European, national or regional initiatives, funding programmes and platforms, in particular with EU-level initiatives such as EOSC^[3],EuroHPC Joint Untedtaking^[4], ESFRI,^[5]AI Factories,^[6] the EU AI, Data and Robotics Partnership,^[7]AI4EOSC^[8] the AI on Demand Platform^[9] and the GenAI4EU Central Hub.

[1] See European Commission’s Group of Chief Scientific Advisors (2024). Successful and timely uptake of artificial intelligence in science in the EU: scientific opinion. Brussels: European Commission.

[2] See Ursula von der Leyen, Political Guidelines For The Next European Commission 2024−2029

[3] Home | European Open Science Cloud - EU Node and the selected projects from the HORIZON-INFRA-2025-01-EOSC-04: Advancing AI-readiness and Machine-Actionability in the EOSC Ecosystem.

[4] Homepage - EuroHPC JU

[5] www.esfri.eu

[6] AI Factories | Shaping Europe’s digital future

[7] Adra Association

[8] AI4EOSC - AI4EOSC

[9] AI-On-Demand Portal - AIoD - AI on Demand