The article already acknowledges that anyone with bad intentions will circumvent any measures like this, and any kind of mandated audio filter would likely just be parsed by people as a connection issue, especially over the phone with a bunch of VoIP compression already. If anything, I think requiring an actual spoken disclaimer that you're talking to an AI with an AI generated voice, like requirements most states have for disclosing calls being recorded, would make more sense.

Rather have a automated "natural voice" because most of the call centers I have delt with have foreign accents that are very difficult to listen to and they miss most of the muinced language of my country.

Why? Talking to a human scammer or an AI scammer -- the ring modulation doesn't change anything. I don't really care to talk to anyone unless it's someone I know from real life, or it's someone in person. For everything else, I literally don't care if it's AI or not -- as long as they're solving the problem for which I called.