"users must provide a verbal consent recording from the voice owner that matches the reference speaker before a voice can be created"
How long is this stored? What could go wrong? :P
Does it detect synthesized consent recordings?
Does it detect synthesized consent recordings?