European Strategy for Artificial Intelligence in Science
09.10.2025
The European Commission presented on 8.10.2025 the strategy “A European Strategy for Artificial Intelligence in Science”, which launches the pilot of the RAISE virtual institute and opens up priority access for scientists to the EuroHPC/AI Factories computing resources, “Data Labs” for working with data and new funding instruments in Horizon Europe 2026–2027.
The document details the roadmap for creating a European “AI in Science” ecosystem based on shared access to computing resources, standardized data and competences. The central element is RAISE — a virtual institute that brings together the EuroHPC/AI Factories (and future AI Gigafactories), the European Open Science Cloud (EOSC) and thematic “Data Labs”.
The latter will serve as “quality mechanisms” for research data: metadata harmonization, dictionary/ontology management, synthetic set generation, legal compliance and readiness for cross-sectoral reuse (including links to EHDS in the health sector).
The strategy envisages networks of excellence and RAISE doctoral programmes, JRC AI Evaluation Hub, as well as systematic updates of “living guidelines” for the safe/responsible use of genSHI. Pilots in materials science (Materials Commons, automated laboratories, foundation models for discovery) and bio/medical sciences demonstrate a focus on rapid translation of results.
At the funding level, reinforcement of “AI in Science” in Horizon Europe 2026–2027 is foreseen, including lab automation and scientific foundation models, as well as new access modes in EuroHPC for academic workloads. For national ecosystems, the document formulates expectations regarding the compatibility of standards, reduction of fragmentation (anti-silo approach), transparent metrics of AI engagement and institutional implementation of ethics/integrity.
The strategy creates a window of opportunity for Ukraine: integration of institutional repositories and collective use centers into European data/computing chains, preparation of joint applications for HE 2026–2027, localization of general AI guidelines and launch of laboratory automation demonstration cases.
