Should callers be told they’re talking to an AI?
Yes — tell them, plainly and in the first breath. It’s what we’d call table stakes for doing this honestly: the caller is giving you their time and possibly their money, and they’re owed the basic fact of what they’re talking to. Beyond ethics, disclosure is simply the stronger business position. A caller who finds out mid-conversation that the “person” was software feels fooled, and fooled customers talk. A caller greeted by a confident, clearly-labeled assistant that immediately helps them just feels… helped.
The trust math
Owners worry disclosure will scare callers off. In practice the fear is backwards. What damages trust is the uncanny moment — the caller growing suspicious, testing the voice, realizing they were meant not to notice. What builds trust is competence plus honesty: the assistant says what it is, handles the request quickly, and offers a human for anything it can’t do. The businesses that benefit most from AI answering are the ones whose setup makes callers think “that was easy,” not “was that a robot pretending?”
What the rules say today
Here is what some of the rules actually say, and whom they bind. The FCC ruled in 2024 that AI-generated voices count as artificial voices under federal robocall law, so calls a business makes with them need the person’s prior consent (written consent, for telemarketing); that’s about outbound calls, not answering the calls you receive. California bars using an automated online account (on a website, an app or a social network) to mislead someone in California about what it is in order to drive a sale; it doesn’t clearly reach phone calls, and a bot that clearly discloses it’s a bot is protected. Texas’s AI law, in effect since January 1, 2026, puts its disclosure duty on government agencies and health care, not on private businesses generally. Rules vary by state, and this isn’t legal advice — ask your own counsel. A business that disclosed from day one has nothing to retrofit and nothing to explain. (It’s the same reason AI-drafted pages on this site carry an AI label.)
How to do it well
Put the disclosure in the greeting, keep it to one short sentence, and pair it immediately with capability: what the assistant can do and how to reach a human. Train it never to claim to be a person, even when asked playfully. And make the handoff real — a caller who says “I want to talk to someone” should get someone, fast. Honesty plus an easy exit is the whole formula.