As Europe looks to unlock the full economic and societal potential of AI, we are facing a simple but important question: will it reflect the continent’s linguistic and cultural diversity?
Right now, many of Europe’s languages remain underrepresented in today’s AI systems. That makes it harder for people to access services, take part in the digital economy, and benefit from new technologies.
To help address this issue, we shared our commitment to support European commerce and culture last July as part of our broader European Digital Commitments. Building on this, Microsoft established the Microsoft Open Innovation Center (MOIC) in Strasbourg, France. Working alongside our AI for Good Lab, MOIC partners with European organizations to help ensure that Europe’s languages and cultural heritage are better represented in AI and digital services. We do this by expanding access to multilingual data and preserving cultural assets so technology can better serve Europe’s citizens, businesses, and values.
Today, we are announcing two new steps to strengthen multilingual AI in Europe, including a partnership with the European Roma Institute for Arts and Culture (ERIAC), a Roma-led organization established by the Council of Europe, and expanded access to multilingual data from GitHub.
The partnership with ERIAC, representing Roma arts and culture organizations across Europe, will support the Amarí Čhib – Romani Language Initiative, a Roma-led effort to address the underrepresentation of the Romani language in AI systems. Despite being spoken by millions of Roma across Europe, it is considered an endangered language characterized by significant dialectal diversity, with varying degrees of mutual intelligibility, limited harmonization efforts, and insufficient digitized content. As a result, Romani speakers risk being excluded from AI-driven systems and opportunities, and the language itself may be used less frequently in everyday, modern contexts. Through the initiative, ERIAC will collect Romani-language text and speech datasets, build an open community-driven digital archive, and prototype AI-powered solutions. The initiative reflects an important principle: language preservation and AI development must go hand in hand with community leadership. By supporting a Roma-led effort, this work helps ensure that cultural context, governance, and priorities are shaped by the communities themselves.
We are also delivering on our commitment to make multilingual data from GitHub and other sources more open and accessible in Europe. GitHub has announced the release of the GitHub Multilingual Repositories Dataset, a metadata dataset designed to help developers find public GitHub repositories where multilingual collaboration may be happening. The dataset is available on GitHub. At a time when many European languages remain underrepresented in digital ecosystems, it serves as a tool for studying language representation in software development. Increasing access to high-quality multilingual datasets is an important step to enabling more inclusive AI systems.
Microsoft efforts here build on work with partners across Europe to reduce cultural and language barriers to AI diffusion. Efforts like the AI for Good Lab’s LINGUA Europe program which brings together universities, researchers, and language communities to develop open datasets, tools, and models for underrepresented languages. Our first cohort spans 16 languages across 10 countries, including Maltese, Luxembourgish, Basque, Romani, and several other regional or minority languages. This effort covers communities of more than 65 million speakers.
And a collaboration with the National Library of the Czech Republic that is helping to digitize and preserve historically significant archives, including the legacy of Václav Havel and the works of the signatories of Charter 77, a human rights manifesto, making them more broadly accessible. These partnerships and others show how digital tools can safeguard cultural heritage and transform collections into resources for research, learning, and civic engagement.
All of these efforts will be a focus of this week’s inaugural Open Innovation Dialogue Hub in partnership with the Council of Europe and GitHub in Strasbourg, where policymakers, researchers, cultural institutions, and technology leaders will explore how open, collaborative approaches can help advance multilingual AI development and preserve Europe’s cultural heritage.
Unlocking the benefits of AI will depend not only on technological progress but also on access to data, strong partnerships, and frameworks that reflect Europe’s values and diversity. Making sure AI systems can support that linguistic diversity will be key to ensuring that everyone can benefit from the opportunities AI creates.