AI for the Many: Anna Korhonen Delivers Keynote at EMNLP 2026

27 October 2026

Our Director, Professor Anna Korhonen, delivered the concluding keynote at EMNLP 2026 in Budapest, exploring how language AI can better serve communities worldwide.

Her talk, “AI for the Many: Starting from the World, Not the Model”, addressed the widening gap between the small number of languages well served by AI and the rest of the world’s languages.

Yet the research highlighted in the keynote shows that more than 1,000 languages already have the foundations for AI. This finding points to opportunities to develop language technologies for many more communities.

The keynote highlighted EQUATE, the World Readiness Index for Language AI, which allows users to explore the readiness of 6,000 languages. It considers available data and models alongside digital infrastructure and socioeconomic conditions, providing a broader picture of where language technologies could be developed and used.

Professor Korhonen called for the needs and priorities of language communities to shape the direction of AI research. Starting with the challenges people face opens up new research questions, grounded in the contexts in which technologies would be used. This means taking account of local resources and infrastructure, and judging success by the benefits AI delivers in practice.

Reflecting on the central message of her keynote, Professor Korhonen said:

“AI that serves the global majority has to be built around the needs and contexts of language communities, on their own terms.”

The keynote drew on research by colleagues at the Language Technology Lab at the University of Cambridge and CHIA, together with partners worldwide.

Find out more

• Explore EQUATE: the World Readiness Index for Language AI
• Read the research preprint: Artificial intelligence is creating a new global linguistic hierarchy
• Read the keynote abstract on the EMNLP 2026 website

Photography: Frame4studio