ElevenLabs is an AI voice platform that clones a voice and dubs video into more than 90 languages while preserving the original tone, timing and emotion. Its Dubbing v2 model launched on May 28, 2026. PR teams can use it to localize an executive statement recorded once, as long as they secure documented consent from the speaker and disclose that the audio was AI-dubbed.
Last updated: October 24, 2026. Product details are as described by ElevenLabs and press coverage of its launch. Check current terms and pricing before use.
In this guide
- What is ElevenLabs and how does it work?
- How does emotion-preserving dubbing work?
- How can PR teams use ElevenLabs for communications?
- What disclosure and consent questions does voice cloning raise?
- What are the limits for PR use?
- What are the capabilities and trade-offs at a glance?
- Frequently asked questions
What is ElevenLabs and how does it work?
ElevenLabs is an artificial intelligence audio company whose products include voice cloning, text-to-speech, speech-to-speech voice conversion and a dubbing model. Its dubbing product translates audio and video across 90+ languages while preserving the emotion, timing, tone and characteristics of each speaker, according to ElevenLabs' own documentation.
ElevenLabs introduced Dubbing v2 in its launch announcement, saying the model conditions directly on the original performance instead of generating audio from a transcript alone.
How does emotion-preserving dubbing work?
Emotion-preserving dubbing works by using the original speaker's delivery as an input, so the translated speech keeps tone, pacing and emphasis, instead of laying a flat translated voiceover on top. ElevenLabs says this makes translated speech feel like the original person said it.
The company also points to cost: it describes professional dubbing as often costing hundreds of dollars per finished minute, with multiple people involved, and says the model automates that process. Treat this as the vendor's positioning, and test output quality yourself in the languages you need.
How can PR teams use ElevenLabs for communications?
A PR team can record an executive statement once and localize it for every market a global company operates in, instead of asking the executive to re-record the message for each language. This is useful for time-sensitive updates across regional press corps.
Illustrative scenario: a company facing a multi-market product recall records one English video statement, runs it through dubbing, and reviews six language versions with native speakers before release. The executive's actual vocal delivery carries across, rather than six different voiceover artists.
What disclosure and consent questions does voice cloning raise?
Voice cloning raises two questions for a communications team: did the person whose voice is cloned give documented consent, and does the published asset disclose that the audio was AI-generated? Get written consent that covers the specific uses, languages and time period, and keep it on file.
Disclosure should be plain and visible, for example a line on the video page or in the caption saying the audio in this language was AI-dubbed from the original recording. Have native speakers and legal counsel review any asset that goes out under an executive's name, and check platform and local-law requirements for synthetic media in each market.
What are the limits for PR use?
Dubbing changes the audio track, not what viewers see, so the speaker's mouth movements stay those of the source language unless a separate lip-sync tool is used. For close-up executive video, test how the result looks and consider using captions or voiceover over b-roll in markets where the mismatch would distract.
Translation quality, names and technical terms also need human review. A mistranslated line in a crisis statement is a bigger risk than a slower release.
What are the capabilities and trade-offs at a glance?
| Use case | Capability applied | Trade-off to plan for |
|---|---|---|
| Multi-market executive statement | Dubbing preserves tone and timing across 90+ languages | Requires documented consent from the cloned speaker |
| Crisis-response video localization | One recording produces several market-ready versions | Needs native-speaker review, and mouth movements stay untranslated |
| Client-facing or long-term voice use | Voice cloning keeps a consistent voice | Needs a clear consent scope, disclosure and periodic review |
Frequently asked questions
What is ElevenLabs Dubbing v2?
A dubbing model ElevenLabs launched on May 28, 2026 that translates audio and video into 90+ languages while carrying over the original speaker's emotion, tone and pacing.
Can a PR team clone an executive's voice?
Technically yes, but only with the executive's documented consent for the specific uses, and with visible disclosure that the audio is AI-generated.
Does AI dubbing change the speaker's lip movements?
Dubbing replaces the audio. Visible mouth movements stay those of the source recording unless a separate lip-sync step is used, so test the result before release.
Should translated statements be reviewed by humans?
Yes. Native speakers should check translation, names and technical terms, especially in crisis or regulated communications.
Conclusion
ElevenLabs addresses a real gap in global communications: a statement recorded once can reach many markets in a voice that sounds like the person speaking. That capability places a direct obligation on the team to secure consent, disclose the AI dubbing and review every language version before it goes out. For related reading on AI and communications work, see the guide to earned media and AI search visibility and 5W's digital PR practice.
