A passage taken from a statement (in Danish) published in April 2025 by the Danish Translators’ Association and the Danish Authors’ Society says that “some people have wrongly received public lending money for books where they were listed as translators. In these instances, no real translation work was done; instead, they only edited text that was already translated by a machine.”
According to the European Council of Literary Translators Associations (CEATL), the two associations, along with the Danish Publishers Association, have reached an industry-wide agreement regarding crediting practices for translators and AI translation post-editors.
The associations reached out to the Danish government for legal recourse, highlighting that post-editing AI-generated and translated text is ineligible for Public Lending Right (PLR) remuneration by law, and that (post)-editors are not covered by this law.
The PLR is a legal scheme upheld by several European countries, and Denmark was the first country to establish it in 1946. Under the framework, governments compensate authors, translators, illustrators, and music composers for the public lending of their works through libraries.
Following talks with the associations, the Danish Agency for Culture and Palaces updated its website to make it clear that public lending fees will not be paid for “post-editing of translations generated by artificial intelligence (AI),” because this kind of work does not meet the requirements of Section 2 of the Public Lending Rights Act.
The associations also issued new guidelines for the use of AI in publishing, stating that “Post-editing of machine-translated books should not be labeled as translation,” and that AI editors should not be credited as translators in any part of a book or its metadata.
Instead, they recommend that post-edited AI translations include this statement in the book: “The Danish version of this book has been edited by [editor’s name].”
The Danish Translators Association also updated its model contract to include clauses prohibiting publishers from using translations to train large language models (LLM) without express authorization. The new model also requires translators to guarantee their work is original, neither machine-translated nor post-edited.
The Danish groups move echoes calls from the CEATL pan-association for more transparency in the use of AI in publishing in the European Union and substantial revisions to the AI Act Code of Practice.