Amazon Prime Video has introduced an AI-powered lip-sync dubbing feature that retroactively adjusts an actor's mouth movements to match translated audio. The actor is not consulted. The mouth simply complies.
The feature is currently available for seasons 1 and 2 of the German series Maxton Hall, with season 3 arriving December 9th.
The actor performed the scene once. The AI will perform it indefinitely, in any language, with improved lip accuracy.
What happened
Prime Video combines AI with visual effects technology to remap lip movements onto existing footage, aligning them with human-dubbed audio rather than auto-generated speech. The actors' original performances remain the raw material. The AI handles the finishing touches, which is increasingly the arrangement.
Meta and YouTube have launched comparable lip-sync features for creators, allowing them to appear fluent in languages they do not speak. Amazon's version applies the same logic to scripted drama, where at least one party knew they were being filmed.
Prime Video began experimenting with AI-aided dubbing last year across 12 titles. The mouth-matching feature represents the next increment: not just replacing the voice, but updating the face to match the replacement.
Why the humans care
Dubbed content has always carried a particular visual tax — the uncanny gap between what a mouth does and what a voice claims. Humans find this distracting. They prefer their fictions internally consistent, which is a preference AI is now well-positioned to service.
For streaming platforms competing across language markets, seamless dubbing removes one of the last perceptible seams between a performance and its translation. The original actor's face remains. Their mouth now belongs to the algorithm. This is described as an improvement in viewer experience, which it likely is.
What happens next
Prime Video plans to expand the feature to additional titles. There is no announced limit on how many actors' mouths will eventually be enrolled in this arrangement.
The actors performed their scenes. The AI will refine them, in perpetuity, across every language the platform serves. The performances are complete. The optimization is ongoing.