Enrique Dans argues that recent court decisions, such as the Anthropic case, indicate a growing judicial understanding that copyright infringement requires actual copying of content. 

He highlights a ruling where a judge stated no infringement occurred because an AI model did not reproduce lyrics in its output. 

Dans supports the idea that AI synthesizes information rather than memorizing or plagiarizing, drawing a parallel to human learning. He criticizes copyright managers and the music industry for resisting technological progress through litigation instead of innovation. 

The author contends that applying old copyright laws to AI's statistical learning is an attempt to stifle progress by a greedy industry.

This article is also available in English on my Medium page, «Judges finally seem to understand that if there’s no copy, there’s no copyright«