Close

Where do we stand in the development of Danish language models?

This question was addressed when the ADD project, in collaboration with IDA, invited different stakeholders to discuss the development of Danish language models.

The stark conclusion at the conference was that tech giants are already too dominant in the market for large language models, and the realistic expectation in Denmark is to develop small, industry-specific or field-defined language models built on top of foreign technological solutions. Not even a Scandinavian language model solution seemed to generate much enthusiasm or appear realistic among the speakers.

One consequence of this dominance is that tech giants are further consolidating control over our digital infrastructure. Additionally, they set the price level for the generative AI solutions they offer, and there is a lack of transparency regarding the data on which the large models are trained.