I don't see why it would matter for an end user answering or calling. I mean, the economics matter (a business can have way more AI voices than hired people to answer calls and send calls). But the experience of the human on the other end is probably ok if the human knows for sure it's an AI they are talking to.

I certainly close all those calls and not bother to interact with them regardless. But in terms of legality I would probably be fine with a restriction and not with an outright ban. Unsure.