Lawyers believe that artificial intelligence is not the owner, author or responsible
The authors object to artificial intelligence (AI) companies that use their images to generate content, but the legal battle will be tough, since both in the United States (EU) and in Europe, the law favors this technology. However, lawyers think that the law can evolve.
In January in the United States, three artists sued Stable Diffusion, Midjourney and DeviantArt. For its part, the Getty photographic agency filed a complaint against Stable Diffusion. The complainants challenge the right of AI companies to use text or images for the “learning” of their programs.
In Europe, a 2019 directive authorizes the right to deep exploration, known as data mining, even on content protected by copyright, if it is publicly accessible. Except, if the owner of the rights expressly opposes it.
“This exception to copyright, designed to allow the development of these technologies, had gone relatively unnoticed,” explained Charles Bouffier of the Racine studio in France.
The difficulty lies in ensuring that the authors’ opposition is respected. For Pierre Pérot, from the August Debouzy studio, located in Paris, it will be difficult to know if a work was used in the apprenticeship phase. Regarding the contents, the legal status is delicate.
Both European and American law only recognize a forgery when a specific work is copied. “Neither one genre nor one style can be protected by copyright,” said Eric Barbry of the Racine studio.
Instead, if the source of an image is clearly recognized, the issue can be raised. In Europe, there is a redoubt that can protect artists copied by AI technologies, as the notion of “free riding” penalizes “looting” the efforts of others.
Another issue raised by the rise of these technologies is the commercial use of content.