The NORA.LLM project - large language models as infrastructure

NORA with partners recently submitted a proposal to the Research Council of Norway under the call National Financing Initiative for Research Infrastructure. The proposal's total budget is 212 million Norwegian kroner (NOK 212M), of which NOK 62M is in-kind contributions from the partners. The goal is to establish Norwegian language models as infrastructure.

We believe it is of paramount importance that Norway establishes a broad national effort for large language models as infrastructure, with focus on our own language and with complete control of the input data. This ensures the preservation of the Norwegian language, Norwegian cultural values and democratic principles, and it may ensure a safe and secure governing model. Many future research projects will strongly rely on such foundational models, and the Norwegian public sector and industry will also build services on top of large language models. This infrastructure, and its coordinated expertise, will strengthen Norway’s position in international research and enhance our capabilities for innovation.

Project partners include Nasjonalbiblioteket, Universitetet i Bergen, Sigma2 AS, Universitetet i Agder, Norsk Regnesentral, Universitetet i Stavanger, Direktoratet for e-helse, SINTEF AS, Oslo Universitetssykehus HF (Kreftregisteret), Universitetet i Oslo, Høgskolen i Østfold, Simula Research Laboratory AS, NORCE Norwegian Research Centre AS, Universitetet i Tromsø - Norges Arktiske Universitet, Arbeids- og Velferdsetaten and Høgskulen på Vestlandet. Norway’s two Centers of Excellence for AI (SFFs), SFF Integreat and SFF Digital Narratives, are supporting the project, so are two Norwegian centers for Innovation in AI (SFIs), SFI Visual Intelligence and SFI Media Futures, as well as the large Cofund programmes DSTrain and LEAD AI

Read the full application with an executive summary here.

We have earlier organized two CuttingEdgeAI seminars on this topic, the talks are highly recommended!

  1. Large Language Models
  2. Investment in infrastructure will determine Norway's intelligent future

We also recommend the podcast NORA forklarer kunstig intelligens med Lilja Øvrelid (in Norwegian, can also be found on Spotify, Apple Podcast and Soundcloud).

I think the best way to think about this is to think about the type of AI systems that have been released so far as being basic infrastructure in themselves ... AI is going to become a common platform and because it's a common platform it needs to be open source, if we want it to be a platform on top of which a whole ecosystem can be
built.
- Yann LeCun in The US Senate hearing on Artificial Intelligence, 19.09.23 [YouTube]

Publisert 7. feb. 2024 06:41 - Sist endret 9. okt. 2025 12:56
Del på e-post