Our feedback about the EU’s AI strategy for the cultural and creative sectors
The European Commission is working on an artificial intelligence (AI) strategy for the cultural and creative sectors.
Autor: Fotograf nieznany
The aim of the strategy is to support innovation, as well as to protect cultural diversity and creators’ rights in the times of rapid development of generative artificial intelligence technologies and services. As part of this process, the Commission launched a Call for Evidence. Our response is based both on extensive research by Centrum Cyfrowe and on years of experience working with museums, libraries, galleries, theatres, and archives on digital transformation.
Developing and implementing a well-rounded Strategy is an opportunity to strengthen not only the creative and cultural sectors, but also European digital sovereignty. For new regulations and guidelines to actually yield such an effect, they must be created in cooperation with those who will be affected – which is why the voices of creators, representatives of cultural institutions, and researchers are crucial here. Above all, it’s the people who have to be prioritized – specifically our right to protect our image and intellectual property, access to diverse culture and knowledge, and living in a sustainable, safe, and pluralistic environment, including the digital realm.
In our response, we call for the EU strategy to be built on four main pillars:
- Protection of creators, their intellectual property, and cultural diversity. Creators must receive fair remuneration for the use of their works to train commercial AI models. We call for mandatory public registries of training data and easily accessible tools allowing creators to submit an effective, binding opt-out from the commercial exploitation of their work. We also propose establishing redistribution systems funded, for example, by an EU digital tax on tech giants, which would balance the disparity between private profits and the costs borne by creators and (often publicly funded) cultural institutions.
- Resolving copyright hurdles in open-source, public interest AI. Cultural heritage institutions often fear legal liability for copyright infringement, which in turn prevents them from fulfilling their public mission in the digital age. We advocate for the introduction of ”Safe Harbour” legal frameworks that would protect institutions acting in good faith. It is also essential to clarify the text and data mining (TDM) exceptions contained in the CDSM Directive and to create a European Public Training Corpus. This would provide access to high-quality, verified, and culturally diverse data – exactly the kind of data that is cared for, researched, digitised and made openly available by cultural and heritage institutions.
- Offsetting increased infrastructure costs for cultural heritage institutions. The digital openness of public cultural and heritage institutions must not be ”penalized” by unrestrained, uncompensated corporate extraction. Large commercial tech companies derive unimaginable profits from AI models trained on, among other things, the aforementioned high-quality digitised collections. While private use, public-interest research, education, and open-source AI development must continue to enjoy frictionless access, these large-scale commercial should contribute proportionally to the underlying public infrastructure.
- Digital sovereignty and the ”Public Money, Public Code” principle. Europe cannot remain dependent on tools and technologies over which it lacks sufficient control. We urge that AI systems, applications, and datasets made with public (EU or national) funding be made available by default under open and free licenses. This builds trust and security through software transparency and auditability, while preventing situations where public institutions pay repeatedly for virtually the same solutions, instead of devoting their funds and energy toward customising them to their needs.
Read our full response on the European Commission’s public consultations platform.