Introducing CLEAR: the Coalition for Language Equity in AI Research

A major shift in the artificial intelligence landscape took root in San Francisco, aimed at making information more accessible to underrepresented communities across the globe. The Coalition for Language Equity in AI Research (CLEAR) held its inaugural meeting, convening nearly 20 leaders from across the technology, academic, and nonprofit sectors. CLEAR is an interdisciplinary group dedicated to a singular, urgent mission: improving the representation of underrepresented languages in AI systems. LILT is proud to support this mission as a founding member, working alongside a diverse group of organizations to improve language representation in AI systems. By bridging the gap between those building the future of technology and the communities often left out of it, the coalition aims to ensure that the AI revolution is truly global and inclusive.

The Imperative for Language Equity

The need for CLEAR is rooted in the stark reality of modern AI development: while systems like Large Language Models (LLMs) are becoming incredibly capable in the small set of languages that predominate in technology, their performance degrades significantly beyond the top 20 languages. With more than 6,000 languages in the world, today’s focus on the top 20 global languages is not enough. The result is a digital divide that threatens to marginalize billions of people, whose languages of choice are inadequately represented in powerful emerging technologies like AI. Technical challenges — such as the scarcity of high-quality, culturally relevant training data and the lack of reliable evaluation standards for so-called “low-resource” languages (LRLs) — have been a barrier for tech companies to expand into these languages.

Beyond the technical hurdles, the human impact of this exclusion is profound. Language is not just a data point; it is a human right and a vessel for cultural heritage, identity, and knowledge. It's a means of economic enfranchisement. When AI fails to support a language, it limits the community's digital presence and their ability to participate equitably in the next generation of the internet. By advocating for language equity, CLEAR seeks to promote community sovereignty, prevent digital marginalization, and ensure that AI provides an equivalent, high-quality experience for all, regardless of the language they speak.

A Cross-Sector Approach to a Global Challenge

Addressing language equity is a highly cross-functional challenge that no single entity can solve alone. The San Francisco meeting reflected this by bringing together a diverse brain trust. Participants include:

• Leading Academics: Experts from Stanford, Berkeley, and UCLA among other premier research institutions.
• Industry Leaders: Representatives from leading LLM providers and AI research labs.
• Global Nonprofits: Thought leaders from NGOs focused on language revitalization and social impact.

This collective expertise is vital for establishing cross-sector alignment on common goals with a focus on ethical community engagement.

The Path Forward

CLEAR’s core philosophy is rooted in the principle of linguistic sovereignty, which recognizes the inherent rights of language communities to determine how their languages are represented, utilized, and developed in AI. This means that language communities are the primary stakeholders and decision-makers regarding the creation, curation, and application of language data and models that pertain to their linguistic heritage. This approach is fundamental to counteracting historical and ongoing patterns of linguistic marginalization and ensuring that AI development is unbiased, equitable, and sustainable.

The inaugural CLEAR meeting established a roadmap for 2026, focusing on three foundational pillars designed to lower barriers and improve language representation in AI.

1. Benchmarking and Leaderboarding: Developing a groundbreaking global survey to assess the current state of language representation and creating authoritative reports on AI quality across different languages.
2. Resource Collection: Building open-access repositories for low-resource language (LRL) datasets and creating modular, open-source infrastructure that can be used by developers worldwide.
3. Strategic Playbooks: Drafting "recipes" and guides for community engagement, ethical data collection, and language expansion strategies to help technology companies and grassroots organizations collaborate more effectively.

A Call to Action

The work of CLEAR is more than a technical initiative; it is a societal imperative. Solutions developed for low-resource languages are, in many ways, solutions for the world’s future, as the majority of the world’s future youth reside in regions where these languages dominate. Ensuring that AI benefits all of humanity requires a sustained, collective effort to protect linguistic diversity as a pillar of digital inclusion.

We invite those in the AI research community, policymakers, and language advocates to join us in this mission. To learn more or to discuss potential alignment with your team, please contact CLEAR at sam.zegas@lilt.com. Together, we can build an AI future where every voice is heard.