For email in the US, I'm not aware of any federal law that requires you to say an AI helped write or send it. The rules that apply are the ones for all commercial email: an honest sender, an honest subject, a real address and a working opt-out. Calls are different: the FCC ruled in 2024 that AI-generated voices count as "artificial" voices under the phone-consumer law, which brings consent rules into play. Beyond the law, there's an honesty rule worth following, and a one-line answer to give if someone asks. This is general information, not legal advice.
Email: you are the sender
When an email goes out from your Gmail, under your name, about your business, you are the sender. Whether you typed every word, used a template or had software draft it doesn't change who's responsible for it. If it says something false, it's your claim. If it ignores an opt-out, it's your violation. That is the right way to think about it: the AI is a tool you use, and you answer for the result. The legal basics are in is cold email legal.
What the law cares about in email is whether the message deceives. A fake name in the From line is a problem. A subject line that misleads is a problem. Saying "I personally watched your video" when no human did, is a problem, because it's a false claim to the reader. A draft written with AI help that says true things about your business isn't.
Calls: where disclosure matters more
Phone is where the rules get stricter. In February 2024, the FCC issued a ruling that calls made with AI-generated voices are treated as artificial voice calls under the TCPA. In practice that means you generally need the called party's prior express consent, and the rules on identifying the caller and honoring opt-outs apply. Texts to leads also need prior consent. If you use AI phone calls for your business, tell the person at the start that it's an AI assistant. A person who feels tricked into talking to a machine hangs up and tells others, and that holds even when the call is legal. Will callers mind an AI answering your phone covers what works on a call.
A few states also have laws about automated "bots" that talk to people to sell something. They vary and change, so if you sell into a particular state at scale, check its rules.
| Channel | What I know of | What to do |
|---|---|---|
| No federal AI-disclosure rule; CAN-SPAM requires honest sender, subject, address, opt-out | Be accurate; keep your own name on it | |
| Phone calls with an AI voice | FCC 2024 ruling: AI voices are "artificial" under the TCPA, so consent rules apply | Get consent; say it's an AI assistant at the start |
| Text messages | Prior consent needed | Only text people who agreed |
| Chat bots on a site | Some states require saying it's a bot | Label it |
The honesty rule
Here's a rule that covers the gaps the law leaves: never claim a human did something that a system did, and never deny it if you're asked directly. You don't have to announce your tools in every email. You do have to avoid inventing a person, a memory or an experience. "I noticed your van is parked outside every Saturday" is fine if you did notice. "I noticed" when nobody noticed is not.
If someone asks, say so plainly
Tomasz owns a sign shop. A prospect replied to his email: "Did you write this yourself?" This is the answer he uses:
"Fair question. I use an AI assistant to help draft and send these, and I read what it sends. It's my business and my account, and anything I promise, I'll do. Happy to talk by phone if that's easier."
That works because it's short, true, and doesn't apologize. Most people don't care that a tool was involved; they care that a real business stands behind the message. Don't dodge the question or reply with a breezy non-answer.
Autonomy and the optional look first
Rocketship's AI workers write, send, reply and book meetings on their own by default, from your own Gmail. If you want to read every email before it goes, "show me first" is a setting you can switch on. The honesty rule applies either way: you chose to use the tool, and your name is on it. How that fits with being a person is covered in hire an AI, not a tool, how replies get handled is in outbound that reads the reply, and what the software does on its own is in what is cold email automation.
What to put in your signature
You don't need a disclosure line. If you'd like one, keep it plain: "Sent by Tomasz Wrona, Wrona Signs" is enough, with your phone number and mailing address under it. Some owners add "I use AI to help with scheduling and follow-ups." That's optional, and it takes the sting out of the question before it's asked. Check the details of what an email must contain in cold email is not spam, and how the writing is judged by readers in do AI-written cold emails go to spam.
When to ask a lawyer
Ask one if you plan to call or text at scale, if you sell into regulated fields such as finance, insurance or health, or if you operate in the EU or Canada, where consent rules are stricter than in the US. Rules on AI are changing quickly, so a short consult costs less than a complaint.
Rocketship finds the buyers, emails them from your own Gmail, follows up, and books the meeting on its own. Plans that get you customers start at $24.99 a month (Launch), and both Launch and Frontier start with a 7-day free trial, card required. Answering your existing number works on Launch, with call minutes bought as you go ($15 for 50), and a dedicated line is $5 a month. See the AI workers.