Given that in practice voice is now just data, it should be revised. But that's expecting the law to match technical reality, which may be overoptimistic...