Denmark

[DK] Extended collective licensing for text and data mining (AI)

IRIS 2024-1:1/11

Terese Foged

Lassen Ricard, law firm

On 1 June 2023 – about two years late – the Danish Parliament passed a bill to implement the Digital Single Market (DSM) Directive in full. This led to the introduction of such things as new rules on text and data mining in sections 11b and 11c of the Danish Copyright Act, equivalent to DSM Directive Articles 3 and 4 (where section 11c corresponds to Article 3 and section 11b to Article 4).

There was no provision in the Danish rules, nor any indication in the legislative history, regarding extended collective licensing in connection with text and data mining, often referred to as the use of Artificial Intelligence (AI).

Rightsholder organisations had lobbied for such licensing, including via a letter of 15 May 2023 which contained a concrete proposal on a specific extended collective licence for reproductions and extractions in connection with text and data mining outside the Copyright Act sections 11b and 11c.

This was reflected in the Ministry of Culture’s letter of 28 May 2023 to the parliamentary Cultural Affairs Committee, stating that the plan immediately following implementation of the DSM Directive was to look, inter alia, at the proposal for an extended collective licence in the field of text and data mining. Thus, a specific extended collective licence of this nature was expected.

On 7 December 2023 the Ministry of Culture sent out for consideration a proposal for amendments to the Danish Copyright Act, with a deadline for comments of 12 January 2024.

The proposal includes no specific extended collective licence for text and data mining, but it stresses that rightsholders’ consent may be necessary; it indicates that the existing general extended collective licence could be used in this connection; and it introduces the possibility of mediation. Thus the proposal states :

“It is further proposed that section 50 paragraph 2 be added to the list in section 52, so that mediation in connection with negotiations under the general extended collective licence in section 50, paragraph 2 be possible.

The Ministry of Culture notes that, for example, in the field of artificial intelligence and text and data mining, there are often large international players, and experience has shown that there may be a need to formalise discussions in negotiations with large tech companies, which mediation can contribute to.

The proposed change means that in connection with general extended collective licensing – e.g. with regard to collective agreements on text and data mining and artificial intelligence (AI) – mediation can be used to facilitate the conclusion of agreements. The parties themselves, i.e. the rightsholders and providers of AI services, can enter into and define agreements, which can then - provided that the conditions are met - be given extended collective licensing effect. This means that such an agreement can define which types of works, etc. are covered. […]

Text and data mining may be associated with the use of material protected by copyright law. This is the case when protected material is copied or content is extracted from a database, for example, when data is normalised during the text and data mining process. In such cases, permission must be obtained from the rightsholders if the exceptions or restrictions in the Copyright Act does not apply.

The exceptions in section 11b and section 11c do not give full access to text and data mining of protected works. The general extended collective licence can be used here to fill the gaps by allowing collective agreements to the extent that use is not permitted under section 11b or section 11c.

It is becoming increasingly common to use automated algorithms (artificial intelligence), whereby a large number of works can be scraped and analysed in a short time. When training models are based on generative AI, taking place during scraping, there is a mass exploitation of works involving a large number of rightsholders. Since the core of the Nordic extended collective licensing model is to ensure agreement on mass exploitation, so that users are ensured easy access to rights clearance, and rightsholders are ensured reasonable payment for the use of their works, the extended collective licence is an option that can be used as a tool in this area.

Like other agreements under the general extended collective licence, agreements on text and data mining under section 50, paragraph 2, have to be approved by the Ministry of Culture."


References


  • 23. maj 2023 - brev fra Kulturministeriet til Folketingets Kulturudvalg med indikationer på kommende ændringer af ophavsretsloven
  • https://www.ft.dk/samling/20222/lovforslag/L125/bilag/8/2712861.pdf
  • 23 May 2023 – letter from the Ministry of Culture to the parliamentary Cultural Affairs Committee with indications on forthcoming amendments to the Copyright Act



  • Høring over forslag til lov om ændring af lov om ophavsret og ændring af lov om kunstneriske uddannelsesinstitutioner under Kulturministeriet
  • https://hoeringsportalen.dk/Hearing/Details/68225
  • 7 December 2023 – hearing about proposal for amendments to the Danish Copyright Act, with deadline for comments of 12 January 2024

This article has been published in IRIS Legal Observations of the European Audiovisual Observatory.