> It is tricking people if it is done to trick people. That's true, but I don't see behaving like a human as "tricking people". Even if the assistant opened th

> It is tricking people if it is done to trick people. That's true, but I don't see behaving like a human as "tricking people". Even if the assistant opened the conversation by identifying itself as an AI, I'd expect it to use all the same mannerisms "uh", "mhmmm" that it did in the call to enable natural conversation. > Would you say the computers in the "Star Trek" shows "behave like a human"? No, but I wouldn't exactly call that "natural interaction" either. Jarvis is a better example. And yes, I do think Jarvis "behaves like a human". > I don't think it's like that at all. Are you impersonating someone else? Gender-bending for kicks? Trying to get better service because you think the person you're calling is sexist or something? In keeping with the analogy, probably closer to "Trying to get better service". An AI callbot that pretends to be a human will be more efficient at its job than one which speaks in a monotone and requires the human on the other end of the call to adjust their mannerisms to compensate for the fact that they're not talking with a human. > And it's not even needed to make the Duplex service work because, as I pointed out before, when a robot calls you to give you money you don't hang up on it! Okay? Depends. If the AI immediately identifies itself as such, how many people are just going to immediately assume it's a spambot and hang up, even when it's not? > people, like you, won't take these concerns seriously What concerns? You've still failed to describe any ill effects that might arise from AI callbots failing to disclose their nature upfront. > You know most hacking is done though what's called "social engineering", eh? You think a law forcing AIs to disclose themselves as AI would do anything against malicious actors who are already breaking the law? (Hacking.)