“If AI systems are trained primarily in majority languages, what space will remain for minority and minoritised languages in the digital future?”
Marta Bainka calls for Digital policies to support language diversity and an ethical A.I. at "Language and AI: Challenges and Opportunities" conference
During the conference “Language and AI: Challenges and Opportunities” held in Barcelona on 5 February 2026, Marta Bainka supported development of artificial intelligence tools to strengthen Europe’s cultural and linguistic diversity and avoid this new technological breakthrough to drive linguistic inequality and discrimination.
Representing the Foundation, Marta Bainka opened the event with a powerful personal reflection. Originally from Silesia, a historic region between Poland, Germany and the Czech Republic, she spoke about growing up in a community whose language, Silesian, still lacks official recognition. As a child, she explained, her education, books and public life were not in her mother tongue. The basic tools to learn and preserve her language were largely absent.
Today, as the world rapidly digitalises, a new tool is shaping our societies: artificial intelligence. For Bainka, the question is clear. “If AI systems are trained primarily in majority languages, what space will remain for minority and minoritised languages in the digital future?”
Bainka believes that “technological development is not neutral. Throughout history, societies have introduced safeguards once the risks of new technologies arose. From seat belts in cars to oxygen masks in airplanes, regulation has not limited freedom, it has protected it.”
AI presents a similar moment of choice. While it offers enormous potential for innovation and progress, it also risks reproducing and amplifying existing inequalities. If AI models are built predominantly from the perspective of dominant cultures and languages, they will inevitably marginalise others.
“If the pre-existing biases of our society are transferred into AI models, and if these tools are built on the worldviews of majority cultures and languages, AI will reproduce and amplify existing inequalities.”
Marta Bainka at “Language and AI : Challenges and Opportunities” conference in Barcelona
Large Language Models (LLMs) require significant volumes of data to function effectively. Where there is insufficient digital production in a given language, these systems will have difficulties to learn, generate content or provide meaningful services for the speakers of languages. The result is a digital ecosystem in which some languages thrive while others are left behind and excluded from circuits of innovation, education and cultural production.
“This raises a fundamental question: do we want some languages to remain on the margins of the digital future?
This is not primarily a technological debate. It is a question of ethics, human rights and the right of all peoples and communities to exist and flourish in the digital age.”
“This raises a fundamental question: do we want some languages to remain on the margins of the digital future?
This is not primarily a technological debate. It is a question of ethics, human rights and the right of all peoples and communities to exist and flourish in the digital age.”
Marta Bainka at “Language and AI : Challenges and Opportunities” conference in Barcelona, 5 February 2026.
. . .
This event is a joint initiative of Fundació Irla , Accent Obert and Coppieters Foundation.
. . .
This event is financially supported by the European Parliament. The European Parliament is not liable for the content of the event nor for the opinions of the speakers.
. . .
Thank you for following our activities over the past few years. We hope our updates have been useful to you. We would like to keep informing you about upcoming events, new publications, summer schools, and job vacancies. Subscribe to our newsletter to hear from us in your inbox.