AI Dubbing in 2026: Voice Actors, Unions and the Real Numbers

The short answer: AI dubbing is now real and in use, but mostly at the edges. Prime Video uses it for catalogue titles that would never have had a human dub, YouTube offers it to every creator in 27 languages, and AI lip sync has reached US cinemas. Every time it has been put on a well-known title with a fan base, the result has been pulled within days, and dubbing actors in Italy, France, Germany, Spain and Mexico have organised to demand consent, pay and labelling before their voices are used to train or drive these systems.

Dubbing means replacing a film or show's original dialogue with a performance recorded in another language. It is a large, mostly invisible craft. Audiences in Italy, France, Germany, Spain and Latin America have grown up with human actors who make foreign films sound native, and many of those actors are household voices in their own countries. Since 2023 streaming platforms and AI companies have been testing whether synthetic voices can do a version of that work faster and cheaper, and the people who do it have pushed back hard.

This is a sourced account of where things stand in October 2026. Every fact below comes from mainstream news, a company's own announcement, a union statement or a named research firm, linked at the end. Where a claim could not be confirmed at source, it is left out. It is written for dubbing and voice actors first, but podcasters who publish video now have AI dubbing available on YouTube by default, so the second half applies to you as well.

What the Platforms Are Actually Doing

Prime Video: a pilot, then a pullback

Amazon announced an AI-aided dubbing pilot on Prime Video in March 2025, covering 12 licensed films and series dubbed into English and Latin American Spanish, among them *El Cid: La Leyenda*, *Mi Mamá Lora* and *Long Lost*. Amazon described it as a hybrid process with human localization experts checking the AI output, and its technology vice president Raf Soltanovich drew the line clearly: "AI-aided dubbing is only available on titles that do not have dubbing support." The pitch, in other words, was reach: titles that would otherwise only ever be subtitled.

Eight months later the same platform showed what happens when that line slips. Over the US Thanksgiving weekend of 2025, Prime Video added AI "beta" English and Spanish dubs to several anime titles, including *Banana Fish* and *No Game, No Life Zero*. Clips spread within hours. Voice actor Daman Mills called the dubs "a massive insult to performers" and cancelled his Prime subscription, and the National Association of Voice Actors (NAVA) called them "AI slop". The licensors were no happier. Kadokawa said it had "not approved an AI dub in any form", and Sentai Filmworks, which had already released a human English dub of *No Game, No Life Zero* in 2017, said the first it heard of the AI versions was when they went live. By early December the dubs were gone, Anime News Network reported.

There is an earlier precedent too. heise reports that in May 2024 Amazon withdrew Latin American Spanish dubs of Korean series after viewers criticised them, without comment. That episode is part of why Mexican dubbing actors, who produce most of Latin America's dubbing, moved early.

MagentaTV: pulled in days

The clearest European example came in February 2025. Deutsche Telekom's streaming service MagentaTV launched the Polish crime series *Morderczynie* ("Murderesses") with a German dub made with AI voices; the credits named the Israeli AI dubbing company Deepdub. Viewers complained that the voices were flat and monotonous, the trade magazine DWDL asked Telekom about it, and the series was removed "for the purpose of quality control and clarification", heise reported on 4 February 2025. Telekom said it would offer the series only in Polish with German subtitles from then on. Deepdub's other work is less contested: in June 2024 Kartoon Studios signed it to dub the Kartoon Channel library into Castilian Spanish, Brazilian Portuguese, Arabic and Italian, Deadline reported.

YouTube: AI dubbing for everyone

The biggest deployment by far is not in film at all. YouTube opened auto dubbing to all users in early 2026, covering 27 languages, with an "Expressive Speech" option in eight of them (English, French, German, Hindi, Indonesian, Italian, Portuguese and Spanish) meant to carry over a creator's tone and energy. YouTube said that in December 2025 it averaged more than 6 million daily viewers who watched at least ten minutes of auto-dubbed content. Viewers get a preferred-language setting; creators can switch auto dubbing off or upload their own human-recorded tracks instead. For podcasters with a video feed this matters directly: unless you change the setting, a synthetic version of your voice may already be speaking to audiences in languages you do not speak.

Netflix: the fight is over training, not dubbing

Netflix has not launched an AI dub track that viewers can choose. Its conflict is about something earlier in the chain. At the start of 2026 its German dubbing contracts gained a clause allowing recordings to be used to train AI systems, without saying whether performers would be paid for it, the voice actors' association VDS told Reuters. A large majority of voice actors refused to agree to the terms, according to VDS chair Anna-Sophia Lumpe. Netflix replied by letter that the concerns were a misunderstanding and, Lumpe said, warned that if the boycott continued some content could be shown in Germany with subtitles instead of a dub. "Our efforts and the efforts of the voice actors are generating a response," she told Reuters. A Netflix spokesperson confirmed the letter and said the company was taking the concerns seriously.

The dispute has since reached well-known shows. Sascha Rotermund, the German voice of Omar Sy, said he would no longer voice Sy in the dispute with Netflix, putting the German version of *Lupin* at risk, Austria's *Der Standard* reported. In September 2026 the VDS filed a complaint asking Germany's Federal Cartel Office to examine whether Netflix is abusing its market power, and backed a separate data protection complaint in Berlin, the German TV trade site fernsehserien.de reported.

AI Lip Sync: Changing the Picture, Not Just the Voice

Traditional dubbing adapts the translation so the new words roughly fit the actor's mouth. AI lip sync does the opposite: it alters the picture so the mouth fits the new words. The most visible example is *Watch the Skies*, a Swedish science fiction film that the company Flawless reworked for English. The original Swedish cast re-recorded their lines in English, and Flawless's software then reshaped their on-screen mouth movements to match, a process it calls "vubbing". AMC Theatres gave the film around 100 screens in the 20 largest US markets from 9 May 2025, Variety reported, and Flawless said the process complied with SAG-AFTRA rules because the actors performed the English themselves.

YouTube is testing the same idea at scale, with a lip sync pilot for selected creators that subtly reanimates the speaker's mouth to match the dubbed audio. Lip sync is the part of AI dubbing that existing contracts handle most permissively. SAG-AFTRA's digital replica guidance for its 2023 film and television agreement lists adjusting a performer's lip, facial or body movement or voice to a foreign language among the uses that do not need fresh consent, and Germany's 2025 AI collective agreement contains a similar carve-out for post-synchronisation into another language, according to a legal analysis on the Kluwer Copyright Blog. Both carve-outs were written with conventional dubbing in mind, which is why unions now watch lip sync closely.

How Big Is the Dubbing Market? Five Estimates, Five Definitions

Search for the size of "the AI dubbing market" and you will find figures that differ by a factor of more than a thousand. That is not because one firm is right and the others are wrong. They are measuring different things. Nimdzi, which tracks the language industry, put the whole language services industry (translation, interpreting, localization and dubbing together) at $71.7 billion in 2024, projecting $75.7 billion for 2025. In a separate October 2024 analysis it put text-to-speech technology, including AI dubbing, at about $5 billion for 2023, with "very high" growth potential.

The narrower reports are much smaller. ResearchAndMarkets valued film dubbing at $2.43 billion in 2019 and forecast $3.60 billion by 2027. Valuates Reports put AI dubbing tools, meaning the software alone, at $783 million in 2023, rising to $1.88 billion by 2030. QYResearch's narrower definition of AI dubbing gives $34.5 million in 2024, rising to $384 million by 2031. Even inside human dubbing the business is consolidating: Nimdzi's 2025 survey found 58.8 percent of language service providers now offer dubbing, voice-over and audio services, and noted that the dubbing company Iyuno was still recovering from the 2023 Hollywood strikes.

The honest reading is this: dubbing is a multi-billion-dollar service inside a $70-billion-plus language industry, and the AI software meant to disrupt it is still a small, fast-growing slice. None of these firms publishes how many human dubbing jobs have actually been lost, and no union has published an audited count either. Treat any article that gives you a precise number for jobs already replaced with suspicion.

Europe's Dubbing Actors Push Back, Country by Country

Italy: a strike, then contract clauses

Italy moved first. From 21 February 2023 about 2,500 dubbing workers, actors, directors and adaptors among them, went on strike over pay, punishing turnaround times and protection from AI; wages had not moved in 15 years, and the strike was extended into March. The unions came out of it pushing contract language that limits a recorded voice to the production it was made for. Sabina Di Marco of the SLC CGIL union told Variety: "We are stipulating clauses that restrict the use of a human voice to a specific product and prohibiting that it be cloned." The veteran dubbing actor Luca Ward put it more bluntly: "You can't sample my voice without my permission!"

The national dubbing actors' association ANAD says those clauses are now in place: a voice hired for one product can only be used for that product, and cloning it needs a separate contract. In 2025 ANAD went further, launching an appeal to defend what it calls "artistic intelligence" and asking for the human voice to be treated in law as biometric data, like a fingerprint, so that misuse could be prosecuted as identity theft. "If this goes on, AI within a few years will blow thousands of jobs in our industry," ANAD president Daniele Giuliani said, in the English translation published by the Italian news agency 9colonne.

France: #TouchePasMaVF and the Stallone voice

In January 2024 the performers' union SFA-CGT and the voice actors' association Les Voix launched a petition "for dubbing created by humans for humans", under the hashtag #TouchePasMaVF ("hands off my French version"). It asks for a clear label when AI is used and for rights holders to be able to refuse AI training on their work, and it had passed 245,000 signatures by December 2025, Journal du Geek reported.

France also produced the case that made the issue personal. In January 2025 the AI voice company ElevenLabs announced it had recreated the voice of Alain Dorval, Sylvester Stallone's French voice for decades, who had died in 2024, for the French release of the film *Armor*. His daughter, the government minister Aurore Bergé, said publicly that she had agreed only to a trial and had not approved the result, CNN reported. A representative of the International Federation of Actors (FIA) called the plan a "provocation". In February 2026 around 4,000 French artists signed an open letter in *Le Parisien*, backed by the performers' rights body Adami, describing AI as a "devouring hydra for artists like us" and warning: "This organised plundering is not hypothetical - it's happening here and now." In the same period, eight French voice actors sent formal legal notices to two US companies that they said had cloned their voices without consent, AFP reported.

Germany: the first AI collective agreement, and a boycott

Germany has the most formal protections and the sharpest current fight. On 10 February 2025 the producers' association Produktionsallianz, the services union ver.di and the actors' union BFFS announced what they described as the first collective agreement on generative AI in German film and television, in force from 1 March 2025. Under it, actors must consent before AI recreates them or clones their voice, and an actor who is fully replaced by AI is still owed pay. The rules were set to run to 30 June 2026, with a review every six months. In June 2025 the BFFS also reached a deal with Netflix requiring written permission before any AI voice replica is used, but deliberately left payment for AI uses open: "there are currently no reference points for appropriate basic remuneration," the union said, as quoted by Reuters.

The VDS, which represents about 600 dubbing and voice actors and was not part of the BFFS deal, rejects that approach. Its members want no AI training on their recordings without consent and payment, and that is the clause behind the 2026 boycott described above. The result is a split worth understanding if you work in Germany: a union deal that covers replicas, and an association fight over training data.

Spain: stop signing until the clause changes

Spanish dubbing actors used the same lever. In January 2024 dubbers across the country agreed to stop working for distributors that would not offer AI protection in the rights assignments they sign before each job, and drafted their own AI clause to send to studios, Spanish dubbing actor and union co-president Sara Gómez Alonso told Variety. The platform of Spanish voice artists' associations and unions, PASAVE, coordinated the clause. The UGT union has warned that AI threatens the work of the roughly 5,000 dubbing professionals in Spain and has called for transparency and collective bargaining before any new contract terms, Forbes España reported.

Beyond Europe: SAG-AFTRA, Mexico and India

In the United States, SAG-AFTRA's 2023 film and television agreement requires clear, informed consent before a performer's digital replica is made or used, with the foreign-language adjustment exception noted above. Its 2024 television animation contracts, ratified by more than 95 percent of voting members, went further for voice work: producers need a performer's consent before using their name to prompt an AI voice, must notify and bargain with the union when they use AI voices in place of actors, and the contract states that "the term 'voice actor' only includes humans", Engadget reported. The negotiating committee co-chairs Bob Bergen and David Jolliffe called it "the first SAG-AFTRA animation voiceover contract with protections against the misuse of artificial intelligence." These protections apply to union productions; much non-union dubbing work for streaming is not covered. Our guide to how SAG-AFTRA's AI terms have worked out three years on goes through them clause by clause.

Mexico, which produces about 65 percent of Latin American dubbing according to figures cited by heise, has moved fastest towards law. On 13 July 2025 dubbing actors, announcers and audio workers marched in Mexico City against voice cloning without consent, under the motto that AI does not replace them, AFP reported. In February 2026 President Claudia Sheinbaum's government sent Congress reforms to the labour and copyright laws that would recognise the human voice as "unique and unrepeatable", forbid cloning it without the owner's consent, and require payment for any AI use. AI dubbing itself would not be banned, heise reported. The dubbing association AMELOC counts 35 active studios and about 1,500 actors in the country.

India shows what is at stake where there is no agreement or law. Ganesh Divekar, president of the Association of Voice Artists, told Scroll that a film with about ten major characters and several minor ones usually needs 30 or 40 dubbing artists for each new language, and that with AI the same job could be done with two male and two female artists supplying source material. "AI needs a command. I do not need a command," he said. The association's general secretary, Amarinder Singh Sodhi, warned that some artists had sold voice samples to technology companies for what seemed a handsome fee and "ended up signing over their voices for life." Internationally, the coalition United Voice Artists, which brings together more than 20 voice actors' guilds, unions and associations from Europe, the United States and Latin America, has lobbied since 2023 for the EU AI Act to cover synthetic voices. Our overview of voice actor unions outside the US and their AI positions covers Equity, ACTRA and others.

How Audiences Have Received AI Dubs

The reception record is consistent. Every AI dub that reached a broad, engaged audience was criticised for the same thing: flat delivery. MagentaTV's viewers called the German voices monotonous. Prime Video's anime dubs were mocked for emotional lines that landed flat, and French coverage described tones that did not fit the scenes. In India, dubbing producer Ashim Samanta told Scroll: "Dubbing requires drama, emotions and expression." YouTube's own response, adding Expressive Speech and planning tools for creators to adjust their dubs, is an acknowledgement of the same gap.

There is a second, quieter pattern. Where AI dubbing has not caused a fuss, it is doing jobs no human was being paid to do: catalogue films that would only ever have had subtitles, creator videos that would otherwise stay in one language, and children's channel libraries. That is the market AI dubbing has won so far. The trouble starts when it moves onto titles that already have, or deserve, a human cast, and onto contracts that quietly turn today's performance into tomorrow's training data. If you want the craft side, our guide to ADR and dubbing sessions explains what a human dubbing actor is actually doing at the microphone that the software has not yet learned.

What It Means for Voice Actors and Podcasters

For dubbing and voice actors, the practical work right now is in the paperwork. The protections that exist, from Rome to Berlin to Los Angeles, rest on the same few ideas, and you can apply them to any contract whether or not you are in a union. Before you sign a dubbing, ADR or voice-over agreement:

1. Search the contract for "artificial intelligence", "machine learning", "training", "synthetic", "digital replica" and "derivative works". These are the words the 2026 Netflix clause and the clauses Spanish dubbers refused turn on.
2. Limit use to the production. Ask for the Italian formula: the recording may be used only for the title and language version you were hired for.
3. Make any replica or training use a separate, written, paid agreement. That is the standard SAG-AFTRA, the German collective agreement and the BFFS deal with Netflix all share.
4. Watch "all media now known or hereafter devised" in perpetuity. On its own it is common; next to a training permission it can cover uses no one has invented yet.
5. Ask whether the platform labels AI voices. France's petition and Mexico's bill both treat disclosure as a basic right.
6. Talk to your union or association before you sign, not after. Collective refusal has moved companies in every country above.

For podcasters, the issue is closer than it looks. If you publish video to YouTube, open YouTube Studio and check the auto dubbing setting on your channel: decide whether you want a synthetic voice speaking for you in other languages, listen to what it produces, or upload your own translated tracks. If you sell host-read ads or license your show, read those contracts for the same training language as above; your voice is the product, and a dub of it is still you. For both groups, our guides to AI voice cloning and consent and the laws that now protect voice and likeness go deeper into the legal side, and the voice-over contracts guide explains usage terms line by line.

The broader picture is not a single fight with a winner. AI dubbing is spreading fastest where nobody was employed before, and it has been rolled back fastest where audiences and performers cared. The contracts being written now in Germany, Italy, Spain, the US and Mexico will decide which of those two directions dominates the next few years, and dubbing actors have shown that saying no together still works.

Frequently asked questions

Is AI dubbing replacing human voice actors?

Not on major releases, as of October 2026. The documented uses are catalogue titles that would never have been dubbed (Prime Video's pilot), creator videos (YouTube) and smaller libraries. Every attempt on a well-known title, including Prime Video's anime dubs and MagentaTV's *Murderesses*, was withdrawn after criticism. The pressure on working actors comes mostly through contract clauses that allow recordings to train AI.

Does Netflix use AI dubbing?

Netflix has not launched AI dub tracks for viewers. Its 2026 dispute in Germany is about a contract clause allowing dubbing recordings to be used for AI training, which the VDS says does not specify payment. Netflix says the concerns stem from a misunderstanding. The VDS took the matter to Germany's cartel office in September 2026.

How big is the AI dubbing market?

It depends on the definition. Valuates Reports puts AI dubbing tools at $783 million in 2023, rising to $1.88 billion by 2030; QYResearch's narrower measure is $34.5 million in 2024. Human film dubbing was $2.43 billion in 2019 by ResearchAndMarkets' count, and Nimdzi sizes the whole language services industry at $71.7 billion in 2024.

What is AI lip sync dubbing?

Instead of fitting the translation to the actor's mouth, AI lip sync alters the picture so the mouth fits the new words. Flawless used it on the Swedish film *Watch the Skies*, with the original cast re-recording their lines in English, for a US release in May 2025. YouTube is piloting a lighter version for creators.

How can a voice actor protect their voice from AI dubbing?

Read every contract for AI training, synthetic voice and digital replica language; limit use to the production you are hired for; make any replica or training use a separate, written, paid agreement; and work through your union or association. These are the terms Italian, German, Spanish and US performers have actually won or are fighting for.

Can I turn off YouTube auto dubbing on my podcast?

Yes. YouTube lets creators switch auto dubbing off in YouTube Studio and upload their own recorded audio tracks in other languages instead. Viewers can also choose whether to hear dubbed audio or the original through a preferred-language setting.

Sources and Further Reading

• Prime Video introduces AI-aided dubbing (Amazon): the 12-title pilot, languages, titles and Raf Soltanovich's statement.
• Prime Video tests AI synchronization of films and series (heise, March 2025): pilot details.
• Answerman: Amazon's AI anime dubs (Anime News Network, December 2025): titles, Daman Mills, NAVA, Kadokawa and Sentai statements.
• Doublage automatique: Prime Video fait marche arrière (Journal du Geek, December 2025): the pullback in France and #TouchePasMaVF's 245,000 signatures.
• "Murderesses": MagentaTV removes series with AI synchronization (heise, 4 February 2025).
• Kartoon Studios and Deepdub dubbing deal (Deadline, June 2024).
• YouTube auto dubbing, explained (YouTube) and YouTube expands auto dubbing to 27 languages (Gulf News): languages, Expressive Speech, the lip sync pilot and 6 million daily viewers.
• German voice actors boycott Netflix over AI clause (Reuters, February 2026): the VDS boycott, Anna-Sophia Lumpe, and the BFFS deal with Netflix.
• Netflix AI dispute with dubbing actors could delay Lupin (Der Standard) and VDS complaint to the Bundeskartellamt (fernsehserien.de, September 2026).
• First collective agreement on AI in the film industry (DGB) and the Produktionsallianz announcement (February 2025).
• From Hollywood to Germany: collective agreements on AI, part 2 (Kluwer Copyright Blog, May 2025): the dubbing carve-outs.
• AMC to bring AI "vubbed" film Watch the Skies to US theaters (AV Club) and Flawless on Watch the Skies (Flawless).
• Nimdzi 100, 2025 and Nimdzi language technology market size (Nimdzi).
• Global film dubbing market forecast (ResearchAndMarkets via Business Wire), AI dubbing tools market (Valuates Reports via PR Newswire) and global AI dubbing market (QYResearch via GII).
• How the SAG-AFTRA strike could help Italy's dubbing workers (Variety, July 2023): Sabina Di Marco and Luca Ward.
• Italian dubbers strike over AI and working conditions (Variety, February 2023) and Dubbing professionals on strike in Italy (MultiLingual).
• Dubbers raise the voice: unfair competition from AI (9colonne, March 2025): ANAD and Daniele Giuliani.
• When Sylvester Stallone's French dubbing artist died, AI recreated his voice (CNN, January 2025) and Sylvester Stallone's French voice and Europe's dubbers (Variety): the Dorval case, the FIA response, and Spain's January 2024 refusal to sign.
• "Devouring hydra": French cinema stars warn AI is plundering industry (AFP via The Local, February 2026).
• UGT warns AI threatens the work of 5,000 dubbing actors (Forbes España).
• SAG-AFTRA ratifies TV animation contracts with AI protections (Engadget, March 2024) and SAG-AFTRA digital replica guidance (SAG-AFTRA).
• AI dubbing controversy: Mexico leads the way (heise, February 2026) and Mexican voice actors protest AI cloning (AFP, July 2025).
• "If AI takes over, we are finished": how Indian voice artists are facing the tech challenge (Scroll, June 2024).
• "Don't steal our voices": dubbing artists confront AI threat (AFP, June 2023): United Voice Artists.