Between the 0 real security and the generational jump in capability fraud now has because of AI, there is no excuse for the continued existence of traditional voice or SMS. We'll never be able to solve compromised devices or social engineering of others, but the idea that anyone/any software in the world can push a message to a device and then immediately communicate with that person is absurd.
At the risk of going too cliché on AI speculation, imagine a hostile AI (or just a non-aligned one under the control of a big nation-state adversary) with nothing but ability to send spoofed texts and calls, and a mission to harm US interests.
It could easily set up thousands of people as unknowing patsies to commit mass terrorist acts, just by giving the AI a small amount of crypto to use in "paying people for gig work" to do tasks. It could start people off with busywork like ferrying harmless packages back and forth, and the ones who proved reliable, eventually would get a much less harmless package to be delivered somewhere sensitive.
That could all be orchestrated by human terrorists of course, but it's so much more effective if they could do it in 100 cities at once without almost any (knowing) human labor, and without any detectable foreign involvement.
