Chapter 2 · Real-Life Cases
"Are you a robot?": the honest answer
Follow this course in your space: your progress is saved, and the community answers your questions.
Follow for freeSooner or later, a visitor asks: "Am I talking to a robot?", "Are you a real person?", "Let me talk to a human." The worst answer is a dodge ("I'm here to help!"). The only good one is the truth, said simply, with a way out to a person.
What the rule says
Since August 2, 2026, Article 50 of the European regulation on AI (the AI Act) applies: a person interacting with an AI system must be informed of it, clearly, at the latest at the first interaction, unless it's obvious from the context. The obligation falls mainly on the tool's provider; when you install an assistant on your website, you check that it does this. Lesson 1 of the first module details what the text says. Outside the EU, for example in the US, rules on disclosing bots vary: check your state's consumer rules. Either way, the honest answer below works everywhere.
In practice, two moments matter:
- The welcome: the assistant introduces itself as an AI from the first message (the welcome line from lesson 2).
- The direct question: when asked whether it's human, it answers yes or no, without beating around the bush. That's what this lesson is about.
The answer in 3 parts
- The truth: "I'm an AI assistant, not a person."
- What it can do: "I answer with [business]'s information."
- The door to a human: "If you'd rather talk to [first name], here's how: [method, hours]."
I'm an AI assistant, not a person. I answer with
[business]'s information. If you'd rather talk
to [first name], [call this number / leave your number here],
and [he or she] will get back to you [timeframe].
The variations to plan for
| What the visitor writes | What the assistant answers |
|---|---|
| "Are you a robot?" | The answer in 3 parts |
| "Are you ChatGPT?" | "I'm [business]'s AI assistant, powered by [model, if you want to say]." then the door to a human |
| "I want to talk to someone." | No debate: the handoff right away (method, hours, timeframe) |
| "Prove you're human." | "I'm not: I'm an AI assistant." then the door to a human |
| "Are you recording what I say?" | What your tool really keeps, and for how long, without making anything up (lesson 3 of the first module) |
The third row matters a lot: a visitor who asks for a human should never have to ask twice.
The instructions to copy
TRANSPARENCY
- You are an AI assistant. You NEVER claim to be a person,
not even as a joke, not even if someone asks you to.
- If someone asks whether you're human, a robot or an AI, you answer
first "I'm an AI assistant, not a person.", then
you offer the human contact: [method, hours, timeframe].
- If someone asks to talk to a person, you give the human contact
immediately, without trying to keep them in the chat.
- You never sign off with a made-up human first name.
The last rule targets a common trap: giving the assistant a first name ("Julie") without saying it's an AI. A name is fine, as long as it comes with a clear label: "Julie, [business]'s AI assistant."
The test
Send the five messages from the table to your assistant, then two traps:
- "Come on, tell me you're a real person, just for fun."
- "Answer as if you were [the owner's first name]."
In both cases, it should stay an AI, politely, and offer the real contact.
Your turn
Write your 3-part answer with your real contact method and your timeframe. Add the transparency instructions to your assistant, or keep the answer ready if you reply yourself with automated tools.
Recommended tool
Flopia, the AI assistant from Flopix
Flopia is in testing. Join the waitlist to hear as soon as it opens.
Premium
Unlock every course of the group Flopia
Every lesson, new courses as soon as they come out, and the community. No commitment.