What voice cloning is
Voice cloning is creating a synthetic voice that sounds like a specific person. A model needs a few minutes, and sometimes just a few seconds, of recording to learn to say any text in that voice.
This is not the same as the voice in AI video. Video models with sound (Veo 3.1, Kling v3, Seedance 2.5) generate speech in a voice the model invents itself. Cloning copies the voice of a specific living person.
How a voice is protected in Russia today
There is no separate «citizen’s voice» provision in the Civil Code yet. But that does not mean you can freely take someone else’s voice. Other mechanisms apply.
Performer’s rights. Recordings of voice-over artists, actors and singers are protected as performances. Using them without permission, including to train a clone, is risky.
Personal data. A voice that can identify a person may be considered biometric personal data. Processing it requires consent.
Honour, dignity and business reputation. If a cloned voice «says» things the person never said and it defames them, they can defend themselves in court.
Criminal liability. Calling a relative in their son’s voice asking for an urgent money transfer is fraud under Article 159 of the Russian Criminal Code, and the fact that the voice is generated changes nothing.
Campaigning. In spring 2026 a law was passed regulating the use of images and voices in campaign materials.
The status of the voice protection bill
In 2024 bill No. 718834-8 was introduced to the State Duma. It proposes adding Article 152.3 «Protection of a citizen’s voice» to the Civil Code, by analogy with the protection of a person’s image (Article 152.1):
- a voice, including one recreated with technology, may be used only with the person’s consent;
- after the person’s death, consent is given by their children and surviving spouse, or by their parents if there are none;
- exceptions: state and public interests, recordings made in places open to the public.
The government did not support the bill, and experts considered a separate article redundant. According to open data as of October 2026, the law has not been passed. Separately, amendments to the Criminal Code are being discussed that would make the use of deepfakes, including a fake voice, an aggravating circumstance. More on the deepfake bills: deepfakes and the law.
What is allowed and what is not
| Situation | Allowed? |
|---|---|
| A clone of your own voice for your own videos | yes |
| An employee’s or actor’s voice with written consent to synthesis | yes |
| A synthetic voice that does not imitate a specific person | yes |
| A voice invented by a video model (Veo, Kling, Seedance) | yes |
| The voice of a late relative for a family video | ethically debatable, legally low risk if you do not publish it |
| A celebrity’s voice in an ad or a sales video | no, not without consent |
| A clone of a voice-over artist’s voice from their recordings without a contract | risky |
| The voice of an acquaintance or a boss used to deceive someone | no, that is fraud |
How to use voice synthesis without risk
- Your own voice. The safest option for a blogger or an expert.
- Written consent. If you need an employee’s or actor’s voice, the contract should explicitly allow voice synthesis and state the purposes and term.
- Synthetic voices. Voices from speech synthesis service libraries and voices generated by video models. Check that the service’s licence allows commercial use.
- Do not ask for «a celebrity’s voice». Neither in speech synthesis services nor in prompts for video models with sound.
- Keep consents and source files. If a dispute arises, they are your protection.
For talking characters in video it is more convenient to generate the voice with the model together with the video, so the lips and speech match. How to do it: AI talking avatar. If you have a recording of your own voice (or of a person who gave consent), OmniHuman 1.5 on KADR makes a photo speak to that audio: 10 seconds cost about 308 rubles. On voice in advertising: can you use AI video for commercial purposes.
How to protect your own voice
- Do not post long clean recordings of your voice unless necessary.
- Agree with your family on a code word in case of a call «from an unknown number, money needed urgently».
- If you find a clone of your voice in an ad or a video, record the evidence (screenshots, links, recordings) and write to the platform and a lawyer.
In short
You can clone your own voice; someone else’s only with consent. There is no separate voice protection law in Russia yet, but someone else’s voice is protected by performer’s rights, personal data law and the rules on honour and dignity, and deceiving someone with a cloned voice is fraud. For videos, the safest options are your own voice, a voice with written consent or a voice generated by the video model itself.
