OpenEuroLLM:
Europe’s Step Towards AI Sovereignty and Innovation
Europe continues to evolve rapidly in digital innovation. This week saw the announcement of OpenEuroLLM, a project in which leading European AI companies and research institutions collaborate to develop open-source language models. Led by Jan Hajič and Peter Sarlin, this initiative aims to build multilingual and transparent AI solutions that align with European values and regulations.
What is Open-Source AI?
Open-source AI refers to artificial intelligence systems that are freely available for use, modification, and distribution. This means that the underlying code, models, and datasets are openly shared, allowing developers and researchers worldwide to collaborate and contribute to the improvement of these technologies. This approach promotes transparency, innovation, and inclusivity.
Why is OpenEuroLLM Important?
Digital Sovereignty
The AI market is currently dominated by large, non-European technology companies, primarily from the United States and China. OpenEuroLLM represents a crucial step towards Europe’s technological independence, reducing reliance on external parties and fostering innovation within the region.
Transparency and Openness
Open-source AI models offer greater control and insight into how language models function. This is essential for ethical AI use and helps businesses and governments make responsible decisions regarding AI technology.
Multilingual Support
A key advantage of OpenEuroLLM is its focus on linguistic and cultural diversity. While many existing AI models are primarily trained in English, this project covers all official EU languages and various European cultures. This ensures that AI remains accessible to all Europeans, regardless of their native language or cultural background.
Wide Application
From businesses and governments to educational institutions and non-profits, the applications of OpenEuroLLM are vast. By making AI technology freely available, organisations across different sectors can benefit from advanced language processing without being dependent on commercial models with closed technology.
A Milestone for AI in Europe
With OpenEuroLLM, Europe demonstrates its commitment to developing responsible, innovative, and inclusive AI technology. This initiative has the potential to strengthen the European digital market, equipping organisations, researchers, and policymakers with the tools to deploy AI fairly and effectively.
Potential Compliance Challenges in Open-Source AI Under the AI Act
The European AI Act, which is already in effect, defines specific roles within the AI value chain—from development to deployment—such as “providers” and “deployers”. Each role comes with its own set of responsibilities and requirements, with providers facing the most stringent obligations. Most organisations using AI for customers, citizens, or employees typically fall under the category of deployers.
However, working with open-source AI models introduces a crucial legal nuance. If an organisation modifies an open-source model to the extent that it becomes a new AI system, it could be classified as a provider under the AI Act. This brings stricter compliance requirements and greater responsibility. The question remains: To what extent are organisations willing to take on these additional obligations, and what strategies will they adopt to manage them effectively?
Investments at the National Level
Beyond European initiatives like OpenEuroLLM, national investments in AI development are also on the rise. For example, the Netherlands is developing GPT-NL, an open language model specifically designed for the Dutch language and context. Supported by the government and universities, this project aims to reduce the country’s reliance on commercial providers. France is also making significant strides, having announced a €10.9 billion investment in AI projects, with a strong focus on innovation and securing a more competitive position in the global AI market. These national efforts highlight that AI is not only a priority at the European level but also a strategic focus for individual countries investing in technologies aligned with their national priorities.
In the coming years, the OpenEuroLLM project is expected to play a pivotal role in shaping AI development within Europe. It marks a step forward towards a more transparent and accessible AI future across the continent.
How PrivacyPerfect Supports Organisations in AI Compliance
At PrivacyPerfect, we closely follow AI developments at the European level. In line with the obligations set out in the AI Act, we offer an AI Pre-Assessments module and an AI Register to help organisations effectively assess and document their AI systems.
AI Pre-Assessment Module
Whether implementing new AI technologies or optimising existing systems, our AI Pre-Assessment Module provides a structured and comprehensive approach to identifying and mitigating AI-related risks—ensuring both compliance and innovation.
AI Register
Our AI Register is a fully customisable, user-friendly solution that enables organisations to document, monitor, and manage their AI systems comprehensively. By providing a clear overview of AI processes, risks, and safeguards, the AI Register simplifies regulatory compliance and ensures organisations remain aligned with global standards.
Would you like to ensure your AI systems comply with the latest European regulations? Get in touch with us to explore how PrivacyPerfect can support your organisation in AI risk management.