CHANSE Call Partner Search Tool

General Information

• Type: Partner looking for project
• Organisation: National Library of Finland
• Country: Finland (FI)

Research area

• Call theme: Crisis - Perspectives from the Humanities
• Keywords:

  data libraries AI

• Brief description of your expertise / expertise you are looking for:

  The recent emergence of commercial AI sets an enormous challenge to research and cultural heritage organisations. How can libraries and other CHOs offer their data to support a transparent development of language models? We need more information not only on juridical and technical aspects but also on the quality of data needed in AI based research. The National Library of Finland is looking for a consortium (and project leader) to answer questions about how libraries (and other cultural heritage organizations) are facing the crisis to provide good quality data for research. Approaches could be defining the prerequisites of using data in research (metadata, collections as data, etc.) and tackling biased data to enable open and ethical AI development. The National Library’s digital materials can offer research resources for different perspectives on crisis The Finnish Web Archive (4.3 billion files and approxiamately 300 terabytes) - online content since 2006, and more material is harvested from websites (e.g., HTM and HTML websites) as well as social media platforms (e.g., Twitter, YouTube, Facebook and TikTok). The National Library preserves Finnish online content and online content concerning Finnish people.  Online material from Finnish domains ending in .fi or .ax, as well as others which have been identified as Finnish-language is harvested annually. This Finnish domain harvest covers approximately 600,000 websites. Thematic harvests relating to a specific topic supplement the annual harvest. Topics have been e.g. Covid-19, Conspiracy Theories and Disinformation, Fake News, Ukraine War and European Refugee Crisis. Online news and newspaper websites are harvested daily, weekly or monthly, according to the updating frequency. Digi.nationallibrary.fi - 26 million digitised pages of history - access to historical newspapers (15 million pages) and journals (seven million pages) published in Finland, ephemera, and other resources digitised by the National Library of Finland.  The most extensive and most used collection in the service consists of newspapers, approximately 60% of the resources. Books are continuously digitised and added to the service too. The openly available newspapers and journals in the service extend up to the end of 1939, as do books whose copyright has expired. The freely available resources include Swedish-language newspapers published up to the end of 1949. Copyrighted resources can be used at the National Library and other legal deposit libraries.

• Brief description of your project / the project you would like to join:

  Ideally, the project increases the readiness of libraries (and other cultural heritage organizations) to react to user centred demands, and to respond to more transparent and data-centric AI development for research. Libraries need to understand researchers’ data requirements to develop data service systems as researchers are relying on data when they develop language models, interpret the past, analyse contemporary public discussion, etc. This requires collaboration and facilitating discussion.

Liisa Näpärä

• Organisation: National Library of Finland
• Position: information specialist
• E-mail: liisa.napara@helsinki.fi