Comment by numpad0

There's nothing that can be done technically. Near perfect voice changing model can be built from 3-5 minutes of conversation on top of a base model, if all the user wants is voice indistinguishable from yours.

But, IMO, the value of mud sludge on a table indistinguishable from sandwiches is tiny. Fake Chanels are 10^2-5 cheaper than the real thing no matter the closeness. Don't listen to people begging you for life to join the counterfeiter ring, they don't make much anyway.