SACEM Doubles Down on Support for France’s AI Training Transparency Bill, Calls Out ‘The Greatest Plundering of Creative and Artistic Works Ever Perpetrated’

SACEM is once again calling for the passage of France’s Darcos Bill (Bill No. 2634), under which it would be up to AI developers to prove “in any civil dispute” that their models weren’t trained on protected works.

The Society of Authors, Composers and Publishers of Music (SACEM) just recently reiterated its support for the Darcos Bill in Paris, where it and many others were on hand for CISAC’s 2026 general assembly.

Unsurprisingly, AI was a major focus during the event, which saw CISAC member-society reps take the stage while holding French-language signs in favor of the legislation. “Creators from all over the world support the Darcos Law,” said signs read, per Google’s translation.

But as the bill is still shelved in France’s National Assembly, it’s SACEM that put out a formal release entreating lawmakers to adopt (or at least advance) the measure.

(On the “still” front, Reporters Without Borders closed out May by describing debate on the bill as “stalled in France’s National Assembly.” In its present June 11th schedule positioning – dead last, that is – the Darcos Bill “has almost no chance of being debated in the chamber,” according to the text.)

Besides the widespread use of a new title (Senator Laure Darcos introduced the legislation), the Darcos Bill remains unchanged from April, when it was unanimously passed in the Senate.

As we reported when SACEM (and SNEP) rallied behind Bill No. 2634 in May, the proposed law would presume that AI developers had utilized protected IP to train their models “unless proven otherwise.”

“Unless proven otherwise, in any civil dispute, the work or object protected by copyright or related rights, within the meaning of this code, is presumed to have been used by the supplier of the artificial intelligence,” the translated document reads.

In this latest show of support, SACEM doubled down on its central position that the unauthorized use of protected works to train gen AI constitutes theft.

“The large-scale use of our works to train AI systems—without authorization or compensation—constitutes the greatest plundering of creative and artistic works ever perpetrated,” SACEM wrote. “And as if that were not enough, our works and the culture of our countries are being overwhelmed and diluted by synthetic content that directly competes with us.”

Taking things a step further, however, the entity touched on the idea that the Darcos Bill’s passage could lay the groundwork for the implementation of similar laws throughout Europe.

“Such a vote would resonate far beyond France’s borders,” SACEM indicated. “There is an urgent need for action on behalf of creators around the world, who know that France—as it has so often done throughout the history of copyright—can once again lead the way and send a clear message: the development of AI must not come at the expense of culture.”

Of course, time will tell whether the Darcos Bill crosses the finish line and whether it spurs different AI crackdowns elsewhere.

But overall, the measure appears to be indicative of the broader regulatory climate for AI developers in Europe, where a German court is now less than two months away from handing down a ruling (originally expected for mid-June) in GEMA’s high-stakes suit against Suno.