2ific
Warning long post.
Generative audio tools are revolutionary & have many uses in sampling , jingles , personalising your events to clients. advertising , remixes.It will also displace many occupations.However there is another side which merits discussion.
Example 1
Voice cloning or mimic with stem extraction or professional studio acapella's
https://youtu.be/Mtg-iTKiXZM?si=fwklqV44TIyN0q1y
Example 2
Generative audio songs with soulful voices.
https://youtu.be/Ck69Fdy1Y44?si=3P3DpaH_qT6wbS8l
Example 3
Deejay software & hardware product demonstration.
https://youtu.be/NopwIVIms5I?si=F7VjaN6thCFipmjU
Ethics & lack of transparency.
Tech giants have been promoting generative audio which use voice models.However professional quality voice models may take up to 100 hours to create & require high spec gpu's.The tech giants enabled free use of online training tools & incentivised users with competitions & referrals.Many services were withdrawn after the developers had enough content or subscribers.
There were some disclaimer forms but the developers must have known that the source of the voice models. The human, the person the Artist were cloned without consent.
The headliner.
Voice cloning used a headliner which was usually a politician or a high profile musician or rapper. This promotion helped normalise unethical behaviour & now many users are conditioned to take someones voice without consent to create , distort , misrepresent.
There are now thousands of unauthorised voice models in the ecosystem.Monetised generative audio tools are accessing those models.The voice was a major component in raising the profile of the tool but the developers are not transparent & declaring it. But they want you to give credit to usage of the tool in your content.
This leads to polarising uncomfortable workflows & scenarios where you have controversial, deceased or unauthorised voices in your songs.
The human the person the Artist
The Artist could be mysterious ,introverted & private.Or a protest act , maverick or renegade. Their recordings , releases & scheduling may have been scarce to prevent over exposure. All those traits & qualities are now meaningless as the voice is now an homogenised preset or prompt.
Any user with access to the tools & the voice can generate an infinite amount of songs & disrupt , distort the aura , vision, brand etc of the act.
Example 4
Unauthorised usage of voices in songs.
https://youtu.be/i-jIEnuN6Ww?si=tMWmKU-E1KnzxlfA
I have more details in the video descriptions & a concise summary here
https://vi-control.net/community/threads/tech-giants-are-normalising-unethical-behaviour-with-ai-voice-cloning-tools.152417/
Posted Thu 04 Jul 24 @ 11:27 am