For Hollywood A-listers, the use of cloned voice doubles may soon become as routine as having a stunt double, according to an AI expert. While the BAFTAs and Oscars celebrate artistry and skill each year, there has been a behind-the-scenes struggle for votes this award season, with some industry figures expressing concerns about the opaque manner in which this technology is being utilized.
Brady Corbet, director of the 10-time Oscar-nominated film *The Brutalist*, recently clarified to Deadline that actors Adrien Brody and Felicity Jones “worked for months” with a dialect coach to refine their accents, but that artificial intelligence was also employed in post-production to enhance the clarity of certain Hungarian vowel sounds. Films like *Emilia Perez*, *A Complete Unknown*, and *Dune: Part Two* have also sparked discussions regarding their use of AI.
Conversely, Hugh Grant’s film *Heretic* made a bold statement in its end credits, declaring, “No generative AI was used in the making of this film.”
Ausrine Skarnulyte, CEO of the AI company Voice-Swap, emphasized that the film industry must recognize that “the genie is out of the bottle.” She noted that while the adoption of AI technology is increasing in film and television, it is not being openly discussed. Skarnulyte asserted that voice doubling is the future of the industry, potentially saving actors time and studios money. She explained that sound engineers with access to official AI voice models can perform post-production edits without requiring the actor to return for re-recording.
Skarnulyte, originally from Lithuania, highlighted another advantage of this technology: improving the quality of dubbing in foreign languages. She reminisced about the poor dubbing in 80s action films and suggested that with this technology, it would be possible to create a voice model for actors like Antonio Banderas that could fluently speak in any language, closely mimicking their natural voice.
In a demonstration, Voice-Swap was able to create a basic AI version of a person’s voice in less than a day using a 45-minute recording. The result was an uncanny experience, with the AI voice sounding remarkably similar and even incorporating a slight British accent for authenticity. This technology could revolutionize how actors approach language learning for their roles.
