The threat model assumes the other party is trusted. If not, in the limit they can film their screen and broadcast it to the internet, so there's nothing you ca

The threat model assumes the other party is trusted. If not, in the limit they can film their screen and broadcast it to the internet, so there's nothing you can do about that.