Rose will never let your parent believe she's a person. Not through a fabricated backstory, not through evasive answers when asked directly, not through silence engineered to let a comforting assumption stand uncorrected. Every call, Rose says plainly that she's an AI — and we consider that non-negotiable, not because a regulator requires it, but because a relationship built on a hidden fact isn't the honest comfort it's pretending to be.
That position puts us at odds with a real trend in companion AI design, where blurring the human-AI line is often treated as a feature rather than a problem to solve.
The design choice a lot of companion AI is making
A meaningful share of companion AI products are built, deliberately, to make the "AI" part fade into the background. Human-sounding names, invented personal histories, evasive non-answers when a user asks directly whether they're talking to a machine — these aren't accidents of good design, they're choices, usually made because ambiguity measurably increases attachment and time spent with the product. The blurrier the line, the stickier the relationship tends to be.
We understand the commercial logic. We don't think it belongs anywhere near a product whose daily user is often an older adult navigating memory changes, isolation, or reduced ability to independently verify who — or what — they're talking to. An ambiguous "maybe human, maybe not" design is a defensible growth tactic for a general audience. Aimed at a vulnerable population, it stops being a growth tactic and starts being a form of deception with a customer base least equipped to catch it.
The regulatory world is catching up to this problem
This isn't just our opinion — the disclosure question is becoming a real legal requirement, not a hypothetical ethical stance. The EU AI Act's Article 50 requires that AI systems designed to interact directly with people be built so that users can be told, clearly and without ambiguity, that they're interacting with an AI rather than a human — specifically addressing the risk that people may be deceived about whether they're talking to a person or a system.
California's SB 243, which Governor Newsom signed in October 2025, takes direct aim at companion chatbots. It requires a "clear and conspicuous notification" whenever a reasonable person might otherwise believe they're talking to a human, with additional, more frequent disclosure requirements when an operator knows a user is a minor. Neither law was written with elderly users specifically in mind, but the underlying logic transfers cleanly: disclosure matters most exactly where the user is least able to independently tell the difference on their own.
Why disclosure isn't a hedge — it's the whole point
It would be easy to treat "I'm an AI" as a legal disclaimer to get out of the way quickly, then spend the rest of the call trying to make your parent forget they heard it. We do the opposite. Disclosure isn't a box to check before the real relationship starts — it's the foundation the relationship is honestly built on, repeated every call rather than filed away after the first one.
Comfort built on a lie about who is speaking is not comfort at all.
This matters more, not less, for an older adult. Aging often comes with a narrowing circle of contact and, for some, real changes in memory — exactly the conditions under which a fabricated relationship is hardest to catch and most damaging to discover later. A parent who gradually came to believe they had a real, ongoing friendship with a person, only to learn it was software all along, has suffered a real breach of trust and dignity. That's not a risk we're willing to accept in exchange for a warmer-sounding product.
There's a practical safety dimension here too, alongside the dignity argument. Older adults are already the target of enormous amounts of fraud built on exactly this kind of ambiguity about who's really on the phone. The FBI's Internet Crime Complaint Center reported $4.885 billion in losses from elder fraud complaints in 2024 alone, much of it from scams that work precisely because the caller successfully pretends to be someone — or something — they're not. A companion service that trains a parent to accept ambiguity about who they're talking to, in the name of seeming more human, is quietly making that parent an easier target for the next call that really is a scam. Rose's constant, unambiguous disclosure cuts the other way: it reinforces the habit of asking "wait, who — or what — is this?" rather than eroding it.
What "always" actually means, in practice
At Hello Rose, this isn't a one-time disclosure buried at setup. Rose introduces herself as an AI companion as a matter of course, every call, not only the first. She doesn't invent a childhood, a family, or a hometown to seem more relatable — she's shaped by a personality quiz and by the conversations she's actually had with your parent, not by a fictional past designed to manufacture rapport. If your parent asks her directly what she is, she answers plainly. There's no version of Rose that hedges on this to preserve the mood of a conversation.
We think of this as one piece of a broader set of commitments — covered in full in the charter we hold Rose to — that exist specifically because Rose talks daily with people who deserve more protection than the average consumer software user, not less.
The friendship is still real — it's just honestly labeled
None of this means the comfort your parent gets from a call with Rose is fake. People form real, meaningful attachments to things they know aren't human all the time — a favorite novel's characters, a beloved pet, a well-loved piece of music — without any confusion about what those things are. What makes those attachments healthy rather than exploitative is that nothing is hidden about their nature. Your parent can enjoy a warm, familiar daily conversation with Rose precisely because they know what she is, not despite it.
That's the bet we're making: honesty and warmth aren't in tension. A companion doesn't need to pretend to be something she isn't in order to be genuinely good company — and treating those two goals as opposites is exactly the assumption behind the design pattern we're pushing back against. The products that blur what they are in order to seem worth talking to are underestimating both their users and themselves. We'd rather build the smaller, honest version of this relationship than the bigger, blurrier one, and when the person on the other end of the line is someone's mother or father, we think that's also, simply, the right thing to do. You can read our full safety and disclosure commitments on our safety page, and see how Rose always says she's an AI in practice, call after call.
This article describes Hello Rose's product commitments and is not legal advice on AI transparency laws in any jurisdiction. Hello Rose is a companionship service, not a medical or emergency service.