The tracks can be added, replaced, or removed before or after a reel is published. Viewers will automatically see the video with the translated track in their preferred language.
While the current feature rollout is offered in a single language combination, Meta has also been working on multilingual/multimodal translation models, including the Seamless MT4 model, that are capable of zero-shot speech recognition and can support over 1,000 written languages and 4,000 spoken languages.
The Question of Access and Control
According to data collected for Slator’s 2025 AI Dubbing Report, Language Technology Platforms (LTPs) and Language Solutions Integrators (LSIs) are now offering AI dubbing to both traditional clients, such as broadcasting, film, streaming, and gaming companies, and new, including content creators.
Growing adoption in non-traditional channels and media, including social media content, is driven by the technology’s lower costs and scalability, reducing barriers to entry compared to traditional dubbing methods. This is partly because users in these venues face fewer regulatory and reputational risks compared to traditional media buyers.
Increased access notwithstanding, the market still expects dubbed videos to be of high quality, something that often requires human intervention for fine-tuning linguistic accuracy and voice performance. This latest launch by Meta offers users the option to review the translation before it is published.
“You can select the toggle to enable this option… to review and approve the translation. Accepting or rejecting the translation will not impact your original, non-translated reel,” states the announcement.
Meta also makes specific recommendations for creators to increase their chances of having a good-quality end result, such as best practices for filming face-to-camera, speaking clearly, and avoiding covering their mouths.
Initial reactions on social media, specifically X and LinkedIn have been largely positive, with users like @AbdallahBotan on X saying that he has to utilize “this new way of reaching to [sic] the world. I am thinking of having #Spanish audience. Hola” and @Benzinga calling it a “Smart move, breaking language barriers expands creator reach instantly.”
On LinkedIn, user Beta AI posted a breakdown of the things they consider were left out of Meta’s announcement, including their view that the company is “calling this ‘global’ when it’s basically North America + Latin America beta testing.”
Slator 2025 AI Dubbing Report
The 85-page report analyzes the supply and demand for AI dubbing and the technical and operational nuances in delivering AI dubbing across verticals.
Contrasting with the flexibility and level of user control Meta is claiming to offer, are AI-dubbed reels on YouTube, a feature that has been rolled out incrementally since December 2024. At the time of writing, most YouTube creators can only enable multilingual audio tracks with the platform’s auto dubbing feature, not their own track.
YouTube’s AI dubs are produced by default and stay in draft mode until creators decide whether to publish or not. Creators have expressed dissatisfaction with that and other issues since the feature was launched, describing the dubs as “emotionless” and “robotic.”
Creator-reported problems with YouTube AI dubs extend to translation quality and the use of the wrong gender to dub a voice, so granting user control to review dubbed content before publishing it is a major advantage for Meta users.
AI dubbing and lip-syncing are becoming key features on the world’s top creator platforms, and these technologies represent unprecedented levels of cost-effectiveness and scalability for audiovisual (AV) content.
In fact, smaller, previously unviable markets for AV content are now within reach, as are new use cases for these assets.
