YouTube Rolls Out Multilingual ‘Expressive Speech’ for AI Dubbing

YouTube has announced that it has expanded its automated AI dubbing feature to include expressive speech capabilities in English, French, German, Hindi, Indonesian, Italian, Portuguese, and Spanish.

The platform’s AI dubbing feature was automatically rolled out to millions of creators in June 2025, and drew heavy criticism from creators and consumers on the “robotic” output of the dubs.

According to Buddhika Kottahachchi, Head of Product at YouTube Dubbing, the new Expressive Speech feature was “developed together with Google DeepMind and […] takes elements like pitch, intonation and energy in the original video and brings [them] into the autodub […] making the dubs seem very natural and authentic.”

As part of the release, YouTube has also rolled out an Automatic Smart Filtering feature that recognizes when a video shouldn’t be dubbed – like music or silent vlogs – to improve the AI dubbing experience. In addition, YouTube confirmed that AI dubs have no impact on the original video’s discoverability, while helping viewers discoverability in other languages.

Despite the improvements, YouTube is also giving control back to users by introducing a Preferred Language setting, enabling viewers to watch videos in the original language, without defaulting to AI-dubbed videos.

The platform rolled out the features in a “commitment to giving creators full agency” over their channels, which also now enables creators to upload their own multi-language audio tracks.

According to Kottahachchi, the average view duration for autodubbed content is approximately 75% of the original language’s duration, indicating significant engagement from non-native speakers.

Kottahachchi confirmed that YouTube plans to roll out its latest Expressive Speech feature in more languages in the near future, as well as its automated lip sync feature, which is being tested by a select pool of creators.