Immigration: no generated clients telling a story

Personal injury for lawyers exists. This page is not that page. Immigration marketing that generates a client telling a story is a likeness problem and a legal-advice problem stacked. Real consented faces, or no faces. There is no third option called "a realistic family at the airport who does not exist."

A deepfake of a private person is not a style. An applicant, a petitioner, a child in a waiting room is a private figure. Law firms already have a lane that works without that replica: know-your-rights explainers, a real attorney on camera, anonymised process — not a synthetic client performing gratitude.

The personal-injury page does not license this

The PI surface is process: what happens in the first week, what evidence to keep, why the early offer arrives. It is not a template for manufacturing injured people, and it is not a template for manufacturing immigrants. Copying the "client story" shape with a generated face inherits two bans at once: private-figure likeness, and a fabricated first-person account.

AI testimonials and the fake review rule is the federal half. A review or testimonial attributed to someone who does not exist, or who never had that matter, is the thing the rule exists to stop. Mechanism is irrelevant. A Seedance extra reading "they got me my green card" is still a fabricated witness. Dramatising a real, granted quote is production. Inventing the client is the violation.

Bar advertising rules still apply. Outcomes, typicality, guaranteed results — the law firm FAQ is blunt: review your jurisdiction before naming any figure. Video does not get a carve-out because a model made the actor. This is general information, not legal advice.

A generated client is a private person

Private means not a public figure. Famous-or-not is the wrong test. If a reasonable person who knows them would say "that is them," you have a likeness problem. If nobody knows them because you invented the face from a "34-year-old mother at arrivals" prompt, you still do not have a client. You have a synthetic extra delivering a legal result they did not live.

Do not seed from a real intake photo "for consistency." That is a replica of a private figure. TikTok's help text is explicit: adult private figures used without permission are not a labelling issue. They are a ban. YouTube, Meta and X will not save you with a different philosophy.

Minors in immigration stories are a hard stop. No generated likeness of anyone under 18. Do not "age them up." Do not put a synthetic child next to a synthetic parent to make the case feel urgent.

A likeness release that covers generation is the document if a real person will be replicated. A 2015 model release for a lobby photograph does not reach new performances. Most immigration clients should never be in that pipeline at all. Film the lawyer. Leave the client out of the generate.

Legal topics need a real lawyer or no face

YouTube's monetisation line on AI personas is narrow and specific: synthetic presenters delivering expert advice on legal issues (and health, finance, politics) as if they were a human expert. AI presenters and sensitive subjects is the map. An invented "immigration attorney" diagnosing a viewer's case is that shape. A named partner, actually licensed, using the tools to explain a public rule is not.

So the production split is:

• Face on camera: a real attorney, or staff who signed a replica grant if you will generate them. Client faces only from consented photographs you will not mutate into new testimony.
• No face: diagrams, type, a process walkthrough that never pretends to be a person. Animated "what a visa category is" is an explainer. Animated "Maria's interview" is a story you do not have the right to tell.

Do not split the difference with a blurred generated extra. Background is not a likeness carve-out, and it is not a testimonials carve-out for a fake result.

What to publish instead of a story

One searchable question per video. What happens at the biometrics appointment. What a Request for Evidence is. What not to say on the first call. That is the same conversion logic the law-firm surface already uses: anxiety-matched, not case-pageant.

If you have a real matter you are allowed to discuss, anonymise it in language — no face, no voice, no invented stand-in "based on" the file. The file is not a character sheet.

PI exists so injury firms can explain process without shouting. That does not mean legal verticals may generate clients. Immigration gets the same discipline, plus a harder likeness line: the people in the stories are private.

FAQ

Can we generate a family at the airport if we never name them?

Not as a client story. An unnamed invented family delivering a visa result is still a fabricated testimonial. An unnamed invented family as anonymous b-roll of "travel" is still a private-figure risk if you seeded a real face. No faces is cleaner.

Does labelling the video AI-generated make a synthetic client legal to run?

No. A label is not a grant and not a real matter. See the private-person post. Permission and truth are separate from disclosure.

Can an AI presenter explain visa categories if we say it is not a lawyer?

Stay in general information, not "here is what you should file." A synthetic expert on legal topics is the YouTube combination to avoid. A real attorney on camera, or no presenter, is the production that does not need that fight.

Why not reuse the personal-injury playbook beat for beat?

Reuse the process beat: calm, specific, no guaranteed outcome. Do not reuse a generated-witness beat. PI's for page is explainers, not synthetic injured people. Immigration should be stricter, not looser.