​At a time when the world is on the cusp of developing technologies that seamlessly integrate into daily life, particularly artificial intelligence, using language as input is a compelling proposition.

Language isn’t just a means of communication; it is a part of culture and identity, something people carry with them wherever they go. Imagine landing in a foreign land, surrounded by unfamiliar sounds, and suddenly hearing a familiar language being spoken. There is an immediate sense of connection, comfort, and belonging. This human response reflects a profound sentiment: Language is one of the forces that anchors human beings culturally, emotionally, and socially.

Linguistic Diversity And The European Dream Of Digital Sovereignty

With 24 official languages, alongside many regional and minority tongues, the European Union stands as a living example of “Unity in Diversity.” Despite such linguistic richness, the European AI landscape is dominated by English and a handful of other majority languages, leaving large parts of Europe linguistically underserved. For example, languages such as Latvian, Irish, and Maltese collectively account for only a small fraction of the datasets used to train the largest commercial models, which means these systems often perform poorly or inconsistently on them.

Having identified this gap, the European Commission launched a new initiative under its Digital Decade Programme 2030 to support the use of more languages in AI. Projects such as the Alliance for Language Technologies European Digital Infrastructure Consortium (ALT-EDIC) and the Language Data Space (LDS) aim to transform how AI is trained across the EU.

Why This Matters: Human and Competitive Edge

With many European languages underrepresented in the current digital landscape, the EU’s approach to embracing and incorporating linguistic diversity into its digital tech plans is a bid to uphold its identity while systematically reducing its dependence on non-EU models. The goal is to foster the EU’s local AI economy and strengthen tech autonomy.

1. Digital Inclusion for All Citizens: When AI tools become operational in native languages, accessibility to the general public is enhanced. This helps close gaps in education, public services, healthcare, legal help, and civic life. It is not just convenient; it is essential for fair access to digital life.
2. Cultural Preservation and Identity: The importance of languages in the history of the human race extends beyond the purview of a set of grammar rules or means of expression to communicate. Languages are cultural identities and a proof of evolution. AI that supports local and indigenous languages helps preserve cultural heritage in the digital world. Ensuring they thrive in technological ecosystems affirms their relevance and vitality.
3. Boosting European Innovation and Competitiveness: Multilingual AI equips European startups and companies with market-tailored tools, reducing reliance on external providers and fostering homegrown innovation. Open models like OpenEuroLLM lower cost barriers, support custom use cases, and help European businesses compete globally on a stronger footing.
4. Strategic Tech Sovereignty: By building its own infrastructure, data, and AI models, Europe depends less on foreign systems that may not comply with local laws, privacy rules, or language requirements. This supports the EU’s goals for digital independence and strength.

A Few Examples From The EU’s Initiatives

ALIA and AINA: Public AI for Spain’s Languages

The Barcelona Supercomputing Center created ALIA and AINA to support Spanish and its co-official languages: Catalan (including Valencian), Basque, and Galician.ALIA provides open language models to encourage innovation and strengthen Europe’s technology and culture. AINA’s goal is to preserve Catalan in the current digital era by providing linguistic and digital resources crucial to processes such as translation, voice assistants, and AI chatbots.

OpenEuroLLM: Multilingual Foundation Models for All EU Languages

Backed by a consortium of research institutions and companies and funded under the Digital Europe Programme, the OpenEuroLLM project is designed to develop a family of open-source Large Language Models (LLMs) covering all official European Union languages, compliant with European AI regulations, and reflecting European values such as transparency and openness.

These models are intended not only for academic use but also for everyday business, public services, and startup innovation, thereby lowering barriers to AI adoption across the EU.

ALT-EDIC and the European Language Data Space

Another cornerstone initiative is the Alliance for Language Technologies EDIC (ALT-EDIC). This multi-country consortium brings together Member States to build a shared infrastructure for language technologies, including data collection, model development, evaluation, and ecosystem support.

Complementing this is the Language Data Space (LDS), a data ecosystem funded under the Digital Europe Programme that facilitates access to multilingual linguistic data for innovation. Together, these efforts address a core challenge: many European languages lack the volume and quality of digital data needed to train robust AI models.

EU Institutional Language Models & Multilingual Services

The European Commission is also developing AI services and models through its institutional channels — for example, efforts to build an EU-wide large language model that supports all official languages and serves public administrations, small and medium-sized enterprises (SMEs), civil society, and academia. These include initiatives like:

• eTranslation (neural machine translation)
• eSummary (automated multilingual summarisation)
• eBriefing (draft document creation) All part of the Digital Europe AI ecosystem and contributing to multilingual public AI services.

Facing the Challenges

No discussion of ambitious AI policy is complete without acknowledging the hurdles:

• Data Scarcity: Many regional languages lack enough high-quality digital data — a core ingredient of powerful AI. Initiatives like ALT-EDIC help, but the work is long and resource-intensive.
• Compute and Energy Costs: Training large, multilingual models requires substantial compute and energy resources. Europe’s infrastructure investment must go hand in hand with sustainability commitments.
• Global Competition: While Europe’s strategy emphasizes openness, values, and inclusion, the global AI landscape is fiercely competitive, with U.S. and Chinese companies investing in vast proprietary systems. Investments such as those under Horizon Europe and Digital Europe are bridging gaps, but the challenge remains significant.
• Coordination and Scale: A shared pan-European approach requires alignment across 27 Member States, each with distinct languages, cultures, priorities, and capacities.

Conclusion: A Future Built for People and Purpose

Europe’s push for linguistic diversity in AI is more than a policy shift influenced by geopolitical changes. It’s a recognition that technology must reflect who we are: culturally, socially,y and democratically. This is a reminder that policies and investments should center not only on innovation but also on the people whose lives are shaped by the languages they speak.

From ALIA’s regional focus to EU-wide initiatives such as OpenEuroLLM and ALT-EDIC, Europe is working toward a future in which AI is inclusive, human-centred, and sovereign, and in which linguistic diversity isn’t an afterthought but a foundational strength.

​

​

europeanpirates.eu/embracing-l…