Across the Nordic and Baltic countries, media organisations and journalists’ unions have issued three joint recommendations aimed at ensuring consent, transparency and effective enforcement when AI companies use journalistic content for commercial purposes.
AI companies use journalistic content to train and develop their systems, often without consent, adequate transparency, or payment. In response, 14 organisations from Denmark, Sweden, Norway, Finland, Iceland, Estonia, Latvia, and Lithuania have published three joint recommendations aimed at strengthening respect for copyright when AI companies use journalistic content for commercial purposes.
Representing media companies, journalists, and rights management organisations, the 14 organisations have come together to establish a shared approach to promoting a fairer market for journalistic content in AI development.
“AI companies are currently using copyrighted content without consent or payment. This threatens not only the financial sustainability of news media but also the ability to maintain a diverse range of independent, editorial media. We need practical solutions that ensure fair terms for the use of journalistic content. Our three recommendations offer concrete ways to address these challenges,” says Karen Rønde, Vice Chair of the Danish Press Publications Collective Management Organization (DPCMO).
Three recommendations for the use of journalistic content in AI
AI needs journalism. The continued development of journalism depends on fair market conditions that make it possible to fund a diverse range of journalism for the benefit of the public.
The 14 organisations have therefore agreed on three specific recommendations:
- Consent: Commercial use of journalistic content in AI requires the consent of rightsholders. AI companies must respect copyright and enter into agreements if they wish to use journalistic content for commercial purposes.
- Transparency: AI companies must provide rightsholders with sufficient information about what copyrighted content is used to train their systems and how copyrighted content is used by those systems. If the necessary information is not made available, a legal presumption with a reversal of the burden of proof should apply.
- Enforcement: Negotiations between press publishers and AI companies have yet to result in agreements. The possibility of establishing an arbitration mechanism or another effective dispute resolution process should therefore be explored. Such a mechanism should help resolve disputes over compensation and related terms without requiring rightsholders to enter into an agreement.
Ongoing legal proceedings highlight the need for clear rules
Several ongoing legal disputes illustrate the need for clear rules governing AI companies’ use of journalistic content.
Earlier this year, DPCMO filed a lawsuit against OpenAI. According to BPCMO, the company used Danish news articles to train ChatGPT and reproduced parts of those articles in violation of copyright law. The lawsuit followed repeated attempts at negotiation and mediation.
The case illustrates the challenges rightsholders face when seeking to enforce their rights against global AI companies.
About the joint position:
The following 14 organisations are behind the joint recommendations:
- Baltic Press Publishers’ Collective Management Organisation
- Danish Media Association
- Danish Press Publishers’ Collective Management Organisation
- Danish Union of Journalists
- Estonian Media Enterprises Association
- Icelandic Union of Journalists
- Norwegian Media Business’ Association
- Norwegian Union of Journalists
- Swedish Magazine Publishers Association
- Swedish Union of Journalists
- Swedish Press Publishers’ Collective Management Organisation
- The Swedish Media Publishers’ Association
- The Finnish Media Federation
- The Union of Journalists in Finland
Read the joint position paper here: Nordic-Baltic Position on Copyright & AI