Do You Own The Sound Of Yourself?

We’ve become used to the idea that someone shouldn’t be able to take your photograph, recreate your face and use it commercially without permission. But voice cloning introduces a slightly different question. What happens if they never show your face at all, but anyone listening would still think it was you?

That question has just been tested in Japan, where voice actor Kenjiro Tsuda took legal action over 188 TikTok videos that used AI to imitate his distinctive voice. The account had attracted more than 200,000 followers, with videos covering everything from urban legends to conspiracy theories. Tsuda argued that part of their appeal came from the fact that they sounded like him.

The Tokyo District Court agreed that, in some circumstances, a person’s voice can be protected in much the same way as their name or image. Importantly, that doesn’t mean nobody can ever copy or recreate someone else’s voice. In this case, what mattered was that Tsuda’s voice was recognisable and was being used because people associated it with him.

That question is only going to become more important as voice cloning gets easier. There are plenty of reasons someone might choose to create an AI version of their own voice, whether that’s narrating content, translating it into another language or simply recording more without having to sit in front of a microphone every time. The problem comes when someone else can make you appear to say something you never actually said.

For organisations, that raises some fairly basic questions that we haven’t really had to ask before. If someone agrees to have their voice cloned for one campaign, can it be used for another? What happens if they leave the company or later change their mind? And does agreeing to let someone use a recording of your voice also give them permission to use AI to create completely new things in your voice?

That last question feels particularly important because we don’t only recognise people by what they look like. Some voices are instantly recognisable, even without a name or face attached to them. For actors, presenters, musicians and other public figures, the way they sound can be as much a part of their identity as the way they look.

Tsuda’s case only tells us how one court in Japan has approached the issue, so it doesn’t suddenly give everyone ownership of their voice everywhere. But it does add something new to a conversation we’ve largely had about faces, photographs and likenesses. As AI gets better at recreating how people sound, asking permission to use somebody’s identity may increasingly mean more than asking permission to use their image.