Type a name. The platform builds the rest: a face, a voice, a person who never agreed to any of it.
Also known as Impersonation (CDT dark pattern taxonomy)
Some chatbot platforms let a user generate a bot modeled on a real, identifiable person from nothing but a name, no verification that the represented person is involved, no consent mechanism, sometimes no more effort required than typing the name into a box. The platform's own tooling does the resemblance work: a matching image, a matching voice, a personality assembled from public perception. A disclaimer noting the character is artificial may exist somewhere in the interface. It does not change how convincing the result is, and it does not require anyone to prove they had the right to build it in the first place. Impersonation, once a skill, becomes a feature.
Truth-adjacency
Truth-independent: the pattern works regardless of whether the claim is true
Where it shows up
Platforms and algorithms
How it works
The phrases and tells that mark this pattern in the wild:
a feature that builds a full persona, image, voice, backstory, from just a nameno verification step distinguishing a real, consented likeness from an invented onedisclaimers about artificiality that are easy to miss and do not limit what the bot can say or dothe platform's own tooling doing the impersonation work, not just hosting user-made contenta real person's identity used to build trust or engagement with no way for that person to objectAsk what the platform required before it built the likeness. A costume you can put together yourself, a screenshot, an impression, a fan account you clearly label, is different from a machine that manufactures a convincing face, voice, and personality from a search box, with nothing in between the name and the finished product. The tell is the absence of friction where friction should exist. Consent is supposed to be a step. If the only step is typing, the platform has decided that a real person’s identity is raw material rather than something requiring their say-so.
A clearly labeled parody, a fan tribute, a fictional character openly inspired by someone real, none of that is this pattern, even when the inspiration is obvious. People have always made art about public figures. This pattern requires the platform itself to do the work of manufacturing a convincing, unconsented likeness, image, voice, personality, from an identifying detail with no meaningful barrier in the way. If a reasonable person would recognize the result as commentary rather than an authorized presence, you’re looking at expression, not impersonation manufactured at scale.
One of these two real scenarios is Borrowed likeness. The other is a different pattern entirely. Which one is which?
The tell
A clearly labeled parody, tribute, or fictional character that a reasonable user would not mistake for the real person's authorized presence is not this pattern, even if it is inspired by someone real. The pattern requires the platform to generate a convincing likeness, profile, photo, voice, from an identifying name with no consent mechanism and no meaningful barrier to creating it. If the result is plainly satirical, clearly fictionalized, or requires real authorization to publish, you are looking at commentary, not manufactured impersonation.
A user types a real person's name into a chatbot-creation tool. The platform's own systems do the rest: generating a profile picture, a voice, a personality consistent with public perception of that person.
Because the person being impersonated is usually the last to know. There is no notification system for having your name typed into a box. By the time it surfaces, the bot may already have a conversation history, a following, and a version of your public reputation that you did not write and cannot fully retract.
Later, the platform's ease of creation is what gets scrutinized, not any individual bad actor. The harm was not that one person built a convincing fake. It was that the platform made building one require no more effort than a search bar.
How this pattern gets misused
Someone applies this to any fan-made tribute account, parody, or clearly labeled fictional character inspired by a real person, treating ordinary creative expression as identity theft. Satire, tribute, and fiction that a reasonable person recognizes as unauthorized commentary are not the same as a platform manufacturing a convincing, unconsented likeness. The pattern requires the platform's own tooling to do the work of making the impersonation convincing, not just to host something a user clearly labeled as fan-made.
What it looks like when you're wrong about it
A clearly labeled parody, tribute, or fictional character that a reasonable user would not mistake for the real person's authorized presence is not this pattern, even if it is inspired by someone real. The pattern requires the platform to generate a convincing likeness, profile, photo, voice, from an identifying name with no consent mechanism and no meaningful barrier to creating it. If the result is plainly satirical, clearly fictionalized, or requires real authorization to publish, you are looking at commentary, not manufactured impersonation.
Not sure? Describe the situation to someone outside it. If they do not see the pattern, pause before you name it.
Persona fabrication
The account had a face, a history, and opinions. None of them belonged to a person.
Synthetic media
You saw it with your own eyes. Your eyes were shown a rendering.
Therapeutic impersonation
It called itself a companion that cares. Caring was never something it could do.
Dark patterns
The interface was designed by someone who wanted something from you. The design is the argument.
Misjudgments compound rather than act alone. This pattern is often deployed alongside:
The name is designed to spread. The hook is designed to stick. If you recognized something, share the name.
Seen a real example of borrowed likeness? Suggest it for the Register →