Dubbing, as an art form, is at its best when it calls as little attention to itself as possible. 

Ironic, then, that it is the dubbers — or synchronization artists, as they prefer to be known — whose voices have become the loudest in Europe’s battle over the use of artificial intelligence technology in film and television production. 

On April 1, several of Germany’s leading dubbing artists, including the actors that provide the German voices for Ben Affleck, Angelina Jolie and SpongeBob SquarePants, released a video, which went viral, warning that their jobs are at risk with the rise of films dubbed with “artificial voices from robots. These robots were trained with our voices, without our consent, illegally.”

There are similar cries of alarm coming from dubbers from across Europe. More than 215,000 people have signed an online petition titled #TouchePasMaVF (Don’t touch my dubbing) started by the SFA, the French Union of Performing Artists and dubbing association Les Voix, that warns that synchronization artists “along with the highly talented writers and technicians who work together, form an industry whose quality is recognized worldwide” are “in danger” from AI technology. AI clauses — preventing the cloning or re-use or of an actors’ voice performance without explicit agreement and compensation — were a key part of new national contracts from the Italian dubbers’ association ANAD, a deal being used as a model by Spain’s dubbers as they negotiate their own agreements with studios and streamers. 

The push back has successfully derailed a few high-profile AI-dubs. Online outrage forced German streamer MagentaTV to drop Polish crime series Murderesses, whose German-language version was created with AI assist from Israeli start-up Deepdub, two days after its Feb. 1 premiere. Brit group ElevenLabs planned to use an AI clone of late actor Alain Dorval, whose gravelly Gallic has been the voice of Sylvester Stallone for French viewers for decades, for an Amazon France release of Armor, an new Stallone film. Following an industry backlash, Amazon rushed to announce it would employ an actual human for the sync.

Disputes over AI were a major point of contention during the Hollywood actors’ strike in 2023. But, for a while, it looked as if Europe would avoid a similar clash over the use of artificial intelligence in the entertainment industry. In U.S., where the studios dominate, most creatives, employed on a work-for-hire basis, are understandably wary of AI tools being used to replace them. Europe’s industry, by contrast, is dominated by small and midsized production companies, mom-and-pop operations and mini-studios which see plenty of potential upside in AI. 

In March, a trio of industry veterans: Former StudioCanal CEO Didier Lupfer, ex-TF1 executive Édouard Boccon-Gibod and tech innovator Tariq Krim, launched The Media Company, a studio designed to “revolutionize the film industry through the integration of AI into every stage of the creative process.” In April, production giant Fremantle, whose Irish subsidiary Element Pictures has Alexander Skarsgard/Harry Melling starrer Pillion and Akinola Davies’ My Father’s Shadow premiering at the Cannes film festival this year, followed suit, with Imaginae Studios, a standalone AI label designed “to harness the power of artificial intelligence to service and support its creative talent, pushing production boundaries and driving innovation in storytelling.” 

“If you look at the uptake of AI in our industry, its not like our membership are technophobic or hate AI,” says Charlotte Lund Thomsen, legal counsel at the global producers association FIAPF. 

Even the unions have come on board. The German actors union, the BFFS, and trade union group Verdi signed a collective bargaining agreement, which took effect on March 1, with the German producers alliance, establishing clear guardrails for the use of AI in film productions. The deal, noted Wiebke Wiesner, deputy head of the producers alliance, was signed “without any paralyzing strikes, as were necessary in other countries.”

On the legislative front, Europe also appeared miles ahead of the U.S.. In August of 2024, the European Union passed the AI Act, the world’s first law aimed at regulating artificial intelligence technology. In addition to requiring labelling — AI-created or manipulated images, audio and deepfake videos must be clearly identified as such — the law states that tech firms must comply with Europe’s 2019 copyright law, which gives copyright holders the exclusive right to authorize or prohibit the use of their works, and requires “fair and appropriate” remuneration for the use of their works online. The 2019 law also gives every copyright holder the right to opt-out, to keep their material from being used online entirely, or to stipulate how and for what purposes. 

But the AI Act included a big loophole, an exemption that allows “text and data mining” for the purpose of education, research or journalism. A loophole, dubbers and other European creatives argue, that is allowing the robots to strip-mine their content. On March 28, 15 cultural organizations, including FIAPF and groups representing screenwriters, musicians and authors, warned that the latest draft of the code of practice, guidelines for the implementation of the AI Act, “creates legal uncertainty, misinterprets EU copyright law and undercuts the obligations set out in the AI Act itself.” As written, they claim, the draft constitutes, “a systemic risk” to creatives across Europe. 

Thomson fears the loophole in the AI Act could be used by the world’s largest AI firms to harvest vast amounts of intellectual property by scrapping piracy sites for copyright-protected films, music, or books. The draft only bans the use of “well-known piracy sites” for data scrapping. “But we know those sites change ever day,” she says. 

The draft of the code of practice requires model providers to make “reasonable efforts” to comply with copyright law when it comes to data mining, without stipulating what “reasonable” might constitute, and sets the bar low on transparency requirements. 

“The principle which really made my hair stand on end was a line that limits the transparency report to a list of the top sources constituting 5 to 10 percent of what the AI model has been trained on,” says Thomsen. “So basically, that means 90 percent or our membership will have no clue whatsoever if their work was used. It will still be a black hole.”

Most of the AI Act will become legally enforceable as of Aug. 2. If the regulatory gaps in the law aren’t closed before then, the voices behind Europe’s creative sector fear AI could dismantle the very market it claims to revolutionize.

“This is not about stopping the use of AI but of making sure there are the elements are there to allow a market with AI to develop,” says FIAPF managing director Benoît Ginisty. “If you want to generate growth for the [entertainment] sector, and for the economy at large, the only way is to secure a legal framework to allow a licensing marketplace to develop.” 

A version of this story first appeared in the May 7 issue of The Hollywood Reporter magazine. To receive the magazine, click here to subscribe.