Watkins, speaking to PCGamesN, elaborated that the text-to-speech always starts with a voice actor: “It’s part of their contract that we use it [AI] for this purpose, and that allows us to do things like our ping system, where it’s capable of saying every single item name, every single location name, and compass directions. That’s how we can get that without needing to have someone come in every time we create a new item for the game.”
So no, they are not “stealing” voices. Their contracts explicity states that they are training the model. So they are getting paid, which in conclusion rules out “stealing”.
Also, from your video:
But to my understanding
Rarely ever good if a sentence starts with “To my understanding”
the AI tools Embark uses to then synthesize the rest of the performance come from models that are trained on millions of other voice actors that have been stolen from in the way that all generative AI models steal from artists.
No, that’s the whole point of models that are trained on a single voice - you do NOT use other voice actors because that would completely muddy the voice. The models are trained on a singular voice to mimic that person perfectly. Using other voices is like asking someone to cook a potato soup for you and then you toss in tomato and paprika.
kate@lemmy.uhhoh.com 2 weeks ago
Stealing? They paid the voice actors to train the model, everyone consented. What do you think stealing is?
ZoteTheMighty@lemmy.zip 2 weeks ago
The job offer would read “Let us use your voice to train an AI to replace you, and if you say no, we’ll hire someone else who will.” Most actors are highly replaceable, so they have very little bargaining power. This is why actors have unions and why they’ve been fighting AI hard.
WindyRebel@lemmy.world 2 weeks ago
I’m admittedly ignorant of all of this, but I’ll post my thoughts anyway:
If the AI is trained on their voice and that AI is available to other studios, the one time payment for that actor’s work is a very very shit deal. They’re out of a job in multiple places or for multiple works now since it can be reused for many things.
I doubt that the contract says it’s only used for that one singular project. Companies are sneaky. However, I have not read said contracts so it’s a magical guess on my part.
Chef6652@lemmy.world 2 weeks ago
Stealing might be a strong word, I agree. But because of this policy, Embark won’t need another recording session out of these actors. Designing the project to never pay the actors that could have worked with them again.
To back this term (“stealing”). I consider that the genAI technology as a whole has been built on the stolen work of the whole world. ChatGPT has been built upon thousands of Github projects without their consent, I suppose Sora, DALL-E, Stable Diffusion, etc have been made in the same way. Training data has been stolen.
Even though it’s a bit far-stretched, using this kind of technology to remove any future human collaboration is stealing to me.
kate@lemmy.uhhoh.com 2 weeks ago
It’s not just a “strong word”, it’s incorrect. At most, these platforms are built on copyright infringement, which is not theft. And in this case, it’s not even copyright infringement since they hired a voice actor
Goun@lemmy.ml 2 weeks ago
I’d say it’s piracy. The same thing some people went to prison for back in the days when dowloading music from p2p.