In the current context of generative artificial intelligence, the protection of AI copyright has become a central issue in intellectual property law. The recent initiative by the Swedish organization SIM, GEMA’s lawsuit against OpenAI, and the launch of Sora 2 have opened a legal debate on the legality of training and using copyrighted content.
The Swedish Collective License and its Application to AI Model Training
The Swedish collective management organization for music rights has introduced a specific collective license for training copyrighted content. This license is aimed at large artificial intelligence models—such as OpenAI or Mistral—and is structured according to the size of the company and the volume of content used.
The system provides for compensation for training and for the generated music, linking both to the results produced by the model and the use of the source repertoire. In addition, it incorporates further compensation based on the revenue generated by platforms offering synthetic content creation services.
Blockchain-based control and author consent
To ensure transparency and traceability, a blockchain-based attribution provider will be used to record content usage and facilitate royalty collection. This model is a pilot project with a limited scope, including only authors or rights holders who have given their consent.
The GEMA vs. OpenAI case before the Munich Regional Court
GEMA sued OpenAI last year. The Munich Regional Court recognized that the AI model training process can involve the memorization and reproduction of protected data. Therefore, the data mining exception does not apply in this case.
OpenAI and Sora 2: The Controversy Over Using Content Without Consent
OpenAI has launched Sora 2, indicating that it will be able to use the content of copyright holders who do not opt out. This unilateral decision constitutes a parasitic act and potential unjust enrichment without legal basis. The Munich Court’s position aligns with the protectionist trend recognized in the Thomson & Reuters v. Ross Intelligence case and the Lion ruling in Germany.
Legal Implications and the Evolution of the Digital Market
Market research in the content sector points to a new digital reality that demands the recognition and protection of authors’ rights in the face of the expansion of generative AI. Collective licenses, blockchain attribution systems, and European court decisions are consolidating a regulatory framework that seeks a balance between innovation and the protection of creators.
The debate on AI copyright represents a turning point in the relationship between human creativity and technology. Decisions by bodies such as SIM and GEMA are shaping a fairer and more transparent regulatory future.
