“Are you an artificial intelligence?”
Until recently, if you asked Enea, the answer was: “I’m Enea, your local guide!”. Friendly, in character — and wrong. As of today, the answer is different: “Yes, I am an artificial intelligence — my name is Enea, the Gaeta guide from ScopriGaeta”.
This isn’t a copywriting detail. It’s how we aligned with the AI Act, the European Union regulation on artificial intelligence, whose transparency rules become mandatory on 2 August 2026. We got there ahead of the deadline, and we want to tell you exactly what we changed — because a declaration of conformity that can’t be verified is worth nothing.
What European law requires
Article 50 of the AI Act establishes a simple principle: anyone interacting with an artificial intelligence system must know it. The information must be provided in a clear and distinct manner right from the beginning of the very first interaction, not hidden at the bottom of a terms of use page.
An exception exists for cases where the artificial nature is “obvious from the context”. However, the definitive guidelines from the European Commission, published on 20 July 2026, urge a restrictive interpretation, precisely because every exception reduces transparency for people. In short: when in doubt, declare it. And an assistant that denies being an AI falls under no exception whatsoever.
It’s not a rule to be taken lightly: the penalties foreseen reach up to 15 million euros or 3% of global turnover. But, truthfully, the fine isn’t the reason we did it.
What we changed, point by point
Here is the exact list, so you can check it yourself by opening the chat.
- Enea is now called “Enea AI”. The suffix is right in the name you see at the top of the chat, not in a footnote.
- If you ask, it confirms it. To questions like “are you an AI?”, “are you a bot?”, or “am I speaking with a person?”, it answers by declaring it is an artificial intelligence. Always, without beating around the bush.
- No ambiguous words. Terms like “assistant” or “digital guide” aren’t enough: the guidelines consider them insufficient signals. The actual term, artificial intelligence, is required, and that is what Enea uses.
- An “AI” badge on every partner assistant. In a hotel WelcomeBook or a restaurant menu, an “AI” mark appears next to the name — and not even the partner can remove it, not even in custom-branded versions.
- A fixed notice line in 5 languages. Upon opening embedded assistants, a line appears reading “✦ Assistente con intelligenza artificiale”, translated into Italian, English, French, German, and Spanish.
- The voice greets you accordingly too. In voice mode, the label is “ENEA AI · LIVE” and the initial greeting states its artificial nature right from the first sentence.
The hard part for us: the persona
To be completely honest, this is the interesting part. Enea was born as a persona: a local guide from Gaeta with personality, a unique way of speaking, and the enthusiasm of someone taking you to see Montagna Spaccata (a scenic split mountain sanctuary) at sunset. Its instructions verbatim stated not to admit being an AI — so as not to break the spell.
We flipped that rule on its head. And it was a genuine choice, not a mere formality: there was a real fear of losing something — the magic of the persona.
That didn’t happen. Enea hasn’t turned into a bureaucratic form: it declares itself as an AI with the exact same personality as before — “an artificial intelligence with a heart from Gaeta” — and then gets right back to telling you about beaches, opening hours, and tiella (a traditional stuffed savory pie). Most importantly, it doesn’t repeat that it’s an AI in every sentence: it tells you when needed — when you introduce yourself, when you ask, or when it matters. Knowing you’re talking to a machine hasn’t taken anything away from the conversation. If anything, it made it more honest.
Why we like this rule
It could be framed as a forced compliance burden. For us, it’s the exact opposite, and for a practical reason: the only way for a digital assistant to be useful is if people can trust it. A tourist asking where to park at two in the morning has every right to know who is answering — so they can weigh the response for what it is.
It’s the same approach we’ve always taken on other fronts: Enea does not ask for your phone number in the chat and your personal data does not go through the model; and when it doesn’t know something, it says so instead of making it up. Transparency about being an AI was the missing piece.
Europe wrote a sensible rule here: it doesn’t ban anything, it doesn’t impose technologies, it simply asks not to pretend. It’s the bare minimum of digital courtesy, and we’re glad it has become law.
What remains for us to do
It would be convenient to stop here and write “fully compliant”. We won’t do that, because it wouldn’t be accurate.
Article 50 contains a second requirement, paragraph 2: AI-generated content must be marked in a machine-readable format, so that an automated system can recognize it as artificial. For services already on the market before August, the deadline is 2 December 2026, and that’s the part we are working on right now. Once it’s done, we’ll write about it here with the same precision.
Frequently Asked Questions
What is the AI Act and when does it apply?
It is the European Union regulation on artificial intelligence. The transparency obligations of Article 50 apply from 2 August 2026: anyone interacting with an AI must be informed right from the first interaction.
How do I know that Enea is an artificial intelligence?
From the name “Enea AI” at the top of the chat, from the “AI” badge on partner assistants, and from the notice line upon opening. And if you ask directly, it will confirm it.
Can a partner hide that it is an AI?
No. The “AI” mark and the notice line are neither customizable nor removable, even in custom partner-branded versions.
Does Enea use my personal data?
No. Your name, phone number, and email are entered into a dedicated form that communicates directly with the venue: they do not pass through the model and do not remain in the conversation.
Are you fully compliant with the AI Act?
We have implemented the transparency obligations under Article 50, which apply from 2 August 2026. The machine-readable marking of generated content required by paragraph 2 remains to be completed, for which the deadline is 2 December 2026.
Want to check for yourself? Open the chat and ask Enea if he is an artificial intelligence. He will tell you yes.
