Artificial intelligence (AI) and language technology have developed rapidly in recent years. The language models used in AI services are trained on vast amounts of text, often without permission from or compensation to the rights holders.

In Norway, there is broad political agreement that we need Norwegian language models. Public and private AI services should reflect the Norwegian and Sámi languages, history, and culture, and there is a need for Norwegian alternatives to international services.

Much is at stake. AI services are already weakening the economic foundation for, among others, translators, illustrators, and those who write and publish educational materials. At the same time, the National Library has documented that language models improve when trained on protected content, such as books and newspapers.

Do Norwegian authors and publishers have a social responsibility to contribute content to Norwegian models if they are compensated?
Would such agreements drain an already vulnerable cultural sector by accelerating a shift where human creativity loses ground?
Perhaps this is a unique opportunity to set a precedent: that training on protected material must be paid for. New income streams for the cultural field could enable the creation and publication of new works of great value to society.

Panel discussion featuring experts in the field:

• Trine Skei Grande, CEO of the Norwegian Publishers Association

• Espen Ytreberg, author and professor of media studies

• Lilja Øvrelid, professor and head of the Language Technology Group at the Department of Informatics, University of Oslo, MediaFutures Work Package 5 member

• Hege Munch Gundersen, CEO of Kopinor

The event is free and open to all. Welcome!