YouTube AI dubbing has been switched on for every creator on the platform since February 4, 2026. No waitlist, no application, no minimum subscriber count. If your channel is eligible, it is already running, and it was probably running before you noticed.
The number that should stop you is not the language count. In December, YouTube averaged more than six million daily viewers who watched at least ten minutes of auto dubbed content. That is not a pilot. That is an audience the size of a national broadcaster, listening to synthetic voices, every day, mostly without thinking about it.
So the obvious question for anybody who charges money for localization is whether the job just evaporated. I sell dubbing and voice work on real productions out of São Paulo, so I have a direct interest in the answer being no. That is exactly why I want to be careful here instead of reassuring.
The honest answer is that YouTube took most of the job. It left one part, the part almost nobody writes about, and that part is worth more than the part it took.
What YouTube AI dubbing gives you for free
Be fair to it first, because it is good, and pretending otherwise is how people end up selling something the market stopped wanting.
Auto dubbing covers 27 languages. On the support side the picture is wider: 34 original languages are supported as a source, and an English video can be dubbed into 23 of them. Most non-English sources primarily get English back, which is a detail that matters if you are Brazilian and assumed this was symmetrical. It is not.
On top of that sits Expressive Speech, available in eight languages: English, French, German, Hindi, Indonesian, Italian, Portuguese and Spanish. It preserves pitch and intonation from the original audio, which is the difference between a translation that sounds like a person and one that sounds like a lift announcement. There is also a Lip Sync feature that alters the speaker's mouth to match the dubbed audio, but be careful how you plan around that one: it is experimental and limited to select channels, not something you can promise a client.
It is on by default for eligible creators. Videos over 120 minutes are out, so are videos with very little speech, very fast speech, an unsupported source language, or a copyright claim.
For a channel talking to the world, this is a genuine gift and it costs nothing. I would not turn it off. Most people asking whether to turn it off are asking the wrong question.
Does YouTube AI dubbing replace a paid dub?
Here is the sentence that decides it, and it is in Google's own documentation rather than in my opinion: you cannot edit automatic dubs.
Read that again with a production head on. Not "editing is limited". Not "editing is coming". You cannot edit them. If the machine mistranslates the one line that carries the joke, or flattens the one word the client spent a week arguing about, your only lever is to change the video's original language setting and ask it to regenerate the whole thing and hope.
Anybody who has delivered paid work knows what that means, because the note always comes back. It is never the whole dub. It is line 14. The client heard line 14 and it lands wrong in their market and they want it changed by Thursday. On a tool with no edit, there is no Thursday. There is only regenerate and pray, or take the dub down.
That is the whole distinction, and it is not about model quality. The YouTube dub and a paid dub can come out of comparable technology and still be different products, because one is a broadcast and the other is a deliverable. A deliverable is something you can be asked to change.
The second difference is quieter and, for me, the bigger one. The voice in an auto dub is not your voice. It is a competent synthetic reading of your words. For a tutorial channel, nobody cares, and they should not. For a person whose face and voice are the product, or for a brand that spent years making a spokesperson recognisable, handing the audio to a generic voice in 23 markets is not localization. It is replacement, quietly, for free, by default.
What I run instead when the dub has to hold up
When a dub is a deliverable rather than a convenience, it goes through ElevenLabs, which is where my trilha, locução and voice cloning already live. Two things about the current state are worth knowing before you plan a job around it, because they moved recently.
Dubbing v2 is the default now and it covers 90 plus languages. It is fully automatic, which means it is fast and it does not let you sit inside a line and change it. To get that, you have to select the older v1 model to reach Dubbing Studio, and ElevenLabs is blunt about the cost: Dubbing Studio is in maintenance mode and receives critical bug fixes only. Find that out now, not on a delivery day. I went through where each piece of this holds up in the longer breakdown of ADR, versioning and narration on paid work.
The part YouTube does not offer at any price is the cloned voice. Not a voice that sounds professional, but the performer's own voice, speaking a language the performer does not speak, still recognisable as them. That is the thing a client is buying when they pay for localization, and I wrote about how it holds up and what it costs in the voice cloning review.
So the split I work to now is simple. Free auto dubbing is for reach: back catalogue, long tail, anything where being understood beats being performed. Paid dubbing is for the pieces where the voice is part of what you are selling, or where somebody has the right to send you a note.
The honest read on what happened here
My contrarian position is that this is not a threat to the craft, and the people treating it as one are defending the wrong thing.
What YouTube automated is the tedious half: get the words into another language, at scale, for free, on ten thousand videos nobody was ever going to pay to localize. That work was not being done before. It was not lost revenue, it was absent revenue, and now six million people a day are getting something they previously got nothing of.
The machine took the daily task. That has been the pattern in every one of these releases and it will keep being the pattern. What it did not take is judgment about which line matters, and the specific human sound of one specific person, because neither of those is a task. A generative tool fed a lazy brief returns a spectacle you have already seen a million times, and a dub with no editor is exactly that: fluent, immediate, and nobody's.
The mistake is to compete with free on the thing free is good at. Nobody wins that. Sell the note. Sell the voice that belongs to somebody.
How to set this up on your own channel today
The practical part, in the order I would do it.
Find out whether it is already on. YouTube Studio, then Content, then Automatic dubbing. It is enabled by default for eligible creators, so check before assuming.
Listen to one dub in a language you speak. Not the language you want to reach. One you can judge. If you speak Portuguese or Spanish, you have Expressive Speech, so you are hearing the good version, not the floor.
Set your original language correctly. This is the one control you have. A wrong source language is the most common cause of a bad dub, and it is the only lever that triggers a clean regeneration.
Decide per video, not per channel. Leave it on for anything informational. Turn it off for anything where your voice is the product, and dub those properly.
Check the pages, not the articles. The state changes and it changes quietly. Google's auto dubbing help page carries the current language list, the eligibility rules and the no-editing statement. YouTube's own announcement is where the rollout date and the viewer numbers come from. For the paid side, the ElevenLabs dubbing guide is where the v1 and v2 difference is documented.
None of this takes an afternoon. It takes about twenty minutes and it is the difference between choosing a default and inheriting one.
Sources and verification
Facts verified September 7, 2026 against two or more independent sources per claim.
- Auto dubbing became available to every creator on February 4, 2026, with no sign-up or waitlist, in 27 languages: YouTube official blog, "Unlocking a global audience with auto dubbing"; corroborated by gHacks, "YouTube Makes Auto-Dubbing Available To All Creators Worldwide" and WinBuzzer, "YouTube Expands AI Auto-dubbing to All Creators in 27 Languages", both February 5, 2026.
- More than six million daily viewers watched at least ten minutes of auto dubbed content in December, and Expressive Speech is available in eight languages: YouTube official blog, same post.
- 34 supported original languages, English dubbing into 23 languages, enabled by default for eligible creators, the 120 minute and copyright ineligibility rules, and the statement that automatic dubs cannot be edited: Google's official YouTube Help page on automatic dubbing.
- Lip Sync is experimental and limited to select channels: Google's official YouTube Help page, corroborated as a pilot in the YouTube blog post.
- ElevenLabs Dubbing v2 is the default model with 90 plus languages, and Dubbing Studio is in maintenance mode receiving critical bug fixes only, reachable by selecting the v1 model: ElevenLabs official product documentation for dubbing.
- No claim is made here about YouTube's internal roadmap, about when Lip Sync leaves testing, or about revenue effects on any specific localization business, because none of that is published.
