Protecting your voice against AI replication is crucial in 2023!
The voice over industry’s concerns over AI continue to grow, and we have now seen multiple cases of AI voice replication without consent.
This includes the original voice of Siri, Bev Standing vs TikTok, and Remie Michelle Clarke’s Revoicer discovery.
So, when it only takes mere minutes of your audio to steal your voice and sell it online, who can you trust to take care of your data?
You can trust Voquent.
We know you’re concerned about your recorded material being used against you in the creation of AI voices, but we have taken as many steps as possible to circumvent this happening on our platform.
From legality to first-person accounts, we are exploring every angle of the AI topic. But what’s all the fuss about, and how are Voquent protecting your data?
How is a voice stolen for AI?
As you can not copyright your voice, only individual performances, there is currently a lack of legal clarity in preventing anyone from creating an AI voice from your recorded material – you don’t actually own it.
Two sets of audio are needed to create an AI voice. First, an extensive audio database is required for the AI to learn how to say any and all combinations of words. After the AI has learned how to speak, as little as 3 seconds of audio is required to replicate how a specific voice sounds and can even attempt to carry forward the tone of a concise sample.
However, the more audio used to influence the sound of an AI voice, the better. Rarely does an AI TTS (artificial intelligence text-to-speech) with this little input sound of professional quality in comparison to real human voices.
How does Voquent protect your voice?
Whilst we can’t solely prevent your voice from being used in AI TTS without permission, several measures are in place to protect voice actors on the Voquent platform.
The demos you upload to Voquent are not downloadable to the users of our site. Clients often request that we send them the demos of the talent they have shortlisted when casting projects. Users can, therefore, only obtain your recorded material when considering you for a project, and no user has the ability to simply download the Voquent database of audio files.
When we send an audio file to one of our clients, we also wipe all ID3 tags, which can contain your personal information in media files. This allows you to be considered for a role without any of your data being shared.
The Benefits of Demo Samples
Typically, when we send a demo sample to a client, we only send one or two. This is another benefit of demo samples; each read is particularly short, so more content is needed to create an effective AI voice. Each sample also represents different tones and styles, so we send the demo samples that match the client’s brief, not your full profile of recorded material.
Even if they did obtain all demo samples, vastly different types of reads are not consistent enough to create a smooth AI voice – One longer consistent performance or multiple reads in the same style make voice replication easier.
Data & Personal Information
Users of the site do not have access to scrape your personal data. This includes the names and addresses of all registered talent. Your client-facing profile does not display your name or address beyond the city you live in. This makes it more difficult for someone to find easily accessible samples of a voice they like from the Voquent site on a different platform.
If you decide to delete your Voquent account, all of your information and audio files will also be removed from the Voquent site. We do not keep a hold of your information. However, if our clients requested that we send them one of your demos while you were still listed on the site, we can’t guarantee that they don’t still have that audio file.
Ultimately, we have taken various measures to protect the Voquent voices. However, when your job is voice acting, it’s not difficult for one of these AI sites to find and collect your audio. If you have voiced even one audiobook, this is far more content than they would need to build an AI voice from your performance. If you have voice over experience of any kind, your voice is already out there to be acquired; we take enough precautions that it would make no sense to get this audio through us, but there are more straightforward means elsewhere.
This is why we need to see more AI regulation in the industry.
Although we can’t protect you from AI TTS theft alone, we will remain vigilant concerning how this technology is changing and how audio is acquired in AI TTS projects.
As our community continues to voice concerns, laws and regulations must catch up with this evolving technology. We’ll be with you each step of the way, keeping our voices updated on new developments while providing as much support and guidance as possible.
Discover the Positive Side of AI Voices
Sometimes we include links to online retail stores such as Amazon. As an Amazon Associate, if you click on a link and make a
purchase, we may receive a small commission at no additional cost to you.