Your voice is indeed a form of authentication to other people that know you. How long until it becomes a valuable token which needs to be protected? I can see

Your voice is indeed a form of authentication to other people that know you. How long until it becomes a valuable token which needs to be protected? I can see various forms of 'man in the middle' attacks taking place with selective conversational interposing allowing an attacker to inject false data into a live conversation between two parties. Imagine a simple example using caller ID spoofing: Attacker calls two parties and uses stolen voices to establish a reason for the call(as both people will know they did not instigate the conversation, both will need to be convinced that the other party did). Now the conversation is started and has context. Now the attacker can steer the conversation to various topics and selectively invert various factual exchanges or negotiations. Afterwards, both parties will be left with a valid memory of the interaction and, unless they confer on who instigated the call, may have little reason to suspect that sections of the exchange were manipulated.