Fish Speech 1.3 is released – Clone Any Voice in 15s
fish.audio
fish.audio
As strong advocates of the open-source community, we've open-sourced Fish Speech 1.2 SFT today and introduced an Auto Reranking system. And we'll be open-sourcing Fish Speech 1.3 soon!
Playground: http://fish.audio GitHub: https://github.com/fishaudio/fish-speech Blog Post: https://blogs.fish.audio/announcing-fish-speech-1-3-30417c28...
How do you deal with the ethics of duplicating someone's voice where it can be misused and/or abused, potentially leading to serious legal, financial, or social consequences for the individual's voice that has been reproduced?
Welcome to the future where you'd have to update that trust and understanding around voice regardless or give that power of misunderstanding to the elite few.
I would appreciate if the question could be answered instead of avoided and misdirected.
It's not going to happen. The only question is whether we're going to have open source solutions, visible to everyone, or whether this kind of thing is going to be available only from shady companies like Cellebrite.
There are methods to manage risk that don't involve turning back the clock. The open source creators /and/ shady companies have responsibility for the tooling they release, just like any other company has responsibility for any potential harm, be it to the environment, infrastructure, or the public that they create.
Yes, there are such things to manage risk, such as insurance. Open source maintainers aren't protected from legal scrutiny or civil suits simply because the software is open source.
Given the amount of false information spreading like wildfire on social media, one would think they would want to protect themselves as much as possible when releasing tooling such as this.