Amazon Prime video plans to use AI to double foreign language emissions and English and Spanish films from Latin America. Company has started a pilot program which uses the “Ai-Aid” dubbing on 12 films and series under license, including titles such as El Cid: The Leyenda, Mi Mamá LoraAnd Long -lost.
Amazon claims that the pilot uses a hybrid approach to double “in which location professionals collaborate with AI to ensure quality control”, and clearly indicates that it will not use its AI-Ai aid process on content that does not already have dubbing support.
I contacted Amazon to find out if the creators of the 12 films and pilot series were involved in the process, but I had not received a response at the time of the publication of this article.
Many moviegoers believe that watching a version doubled by a film or a series in a foreign language undermines art. Since the performance of an actor is a combination of movement, speech and accent, it is important to discover all of this, even if you need subtitles to understand what is said. If the dubbing of the AI could preserve 100% of this performance, while converting it into a different language, it could redefine what a double film means.
On the other hand, the dubbing of the AI threatens the livelihoods of actors in the professional voice. In 2023, The actors of the voice sounded the alarm Via the National Association of Actors of La Voix (Nava). He Advice issued For actors in the voice, tell them never to grant summary rights to a client and to contact their union or a lawyer if they suspect that the contract tries to take its rights.
Among their concerns, studios could use AI to edit dialogue lines, in fact obtaining new performances from the actors without bringing them back to the recording studio (or paying them to do so).
Amazon is not the first company to use dubbing based on AI. In 2023, Spotify began a tool based on Openai technology that allowed it Clone the voices of its podcast hosts And nicknames them in other languages.
This technology continued to improve at a dramatic rate. In 2024, Openai boasted that he Need only 15 seconds of audio sample To create a clone of someone’s voice. A few months later, Microsoft – who invested massively in Openai – revealed that his own model of AI of Pointe Vall -E 2 was 2 was Too dangerous to releaseBased on its realism, which has aroused fears of improper use.