Exploring Generative AI and Copyright in Denmark

Artificial intelligence is not a new thing, but its development has taken huge steps over recent years, with a tremendous amount of new and much more capable tools. It is increasingly important to have knowledge of the attendant legal issues and the relevant legislation, from both a Danish and European perspective.

It has been suggested that the use of copyright-protected material without consent in the training of AI tools is unlawful. If a user prompts the tool to create a product based on such illegal training, the output would also be illegal.

Things may change if new legislation regarding data mining comes into force. No litigation such as the New York Times’ litigation against OpenAi for using its articles is currently pending in Denmark, but the Right Holders Alliance has taken the initiative to stop the use of various tools.

AI regulation contains minimal content on copyright, but activity on the legislative side is expected. However, it is very difficult for legislators to understand the technology, and legislation takes time. Also, the AI environment is evolving at an incredible pace, with new tools arising almost on a daily basis.

Court cases will be decisive in Denmark and in the US, and from the European Court of Justice in particular, as that is where details are likely to be dealt with first. There are expected to be difficulties in compelling huge tech companies to disclose what is actually going on in their algorithms.

Speedy procedures like injunctions may not be effective as it still needs to be proved that the training was illegal. In addition, AI injunctions may go on for weeks and weeks.

It may be practical for certain regulatory bodies to approve AI tools, especially in cross-border scenarios. The EU Commission is increasingly stepping in, not necessarily to decide cases, but to connect the national authorities and to make sure that information from on country reaches another country. It is also active in issuing sanctions.