The use of works, articles, images and cultural content to train artificial intelligence systems raises a sensitive question: how to encourage innovation without dispossessing creators of their work?
A conflict between innovation and remuneration
Generative artificial intelligence relies on considerable masses of data. Texts, images, videos, sounds, archives and content published online can be used to train models capable of producing new content. For technology companies, this raw material is essential to the performance of the tools.
For authors, journalists, photographers, publishers, musicians and artists, the question is different: if their works serve to create value, should they be informed, remunerated or be able to refuse? This question is becoming central in the cultural debate.
Local media also concerned
The subject does not only concern large cultural groups. Local media, news sites, independent photographers and small editorial offices produce original content every day. If this content is sucked up without a clear framework, the value of the editorial work can be weakened.
For a news article site, the issue is very concrete. Writing, proofing, prioritizing and providing context costs time. If AI takes over this content without economic return, the press model becomes even more strained.
Towards a right to transparency?
The compromise could involve more transparency: knowing what data is used, allowing rights holders to oppose certain uses, organizing collective licenses or providing remuneration when the content is used to drive commercial tools.
The article must remain educational. It is not a question of presenting AI as an absolute threat, nor as a magic solution. The real debate concerns the sharing of value. Technology can be useful, but it must not erase those who produce information, creation and cultural memory.
