OpenEuroLLM: Europe Unites for Sovereign, Multilingual AI
Twenty leading European stakeholders have initiated OpenEuroLLM, an ambitious open-source project aimed at developing ethical, competitive multilingual AI models to bolster Europe's digital sovereignty.
Europe Unites for the Future of AI: The OpenEuroLLM Project
In a significant announcement in artificial intelligence (AI), 20 elite European research institutions, businesses, and EuroHPC centers have joined forces to launch an unprecedented initiative: OpenEuroLLM. Coordinated by Jan Hajič and co-directed by Peter Sarlin, this project aims to develop a series of high-performance multilingual foundational linguistic models in open-source.
Addressing the Challenge of Artificial Intelligence at the European Level
Looking towards the future, OpenEuroLLM is a key step in ensuring Europe’s competitiveness and digital sovereignty. The models being developed will democratize access to high-quality AI technologies while boosting the ability of European businesses to excel and innovate globally.
Transparent and Regulated Models
The models will be designed within Europe’s robust regulatory framework, ensuring compliance with its values. They will benefit from collaborations with open-source communities and open science, as well as industry experts through the project’s strategic partnership council. These synergies will meet the specific needs of both the industry and public sector.
Preserving Linguistic and Cultural Diversity
OpenEuroLLM is committed to preserving linguistic and cultural diversity in AI. The multilingual models will provide European businesses the opportunity to develop superior quality products and services in the AI era.
The project, awarded by the Strategic Technologies for Europe Platform (STEP), is supported by prior European projects and the combined experiences of the consortium partners. It will commence its activities on February 1, 2025, with funding from the European Commission under the Digital Europe Program.