Skip to content
Updated for 2026

A name in your prompt — even a common first name in a line of dialogue — can trigger the video model's real-person protections, which treat any name as a possible reference to someone real and decline the render. It's nondeterministic: the same prompt can pass once and fail the next time. Describe characters by appearance instead.

Why Does Using a Real Person's Name Break My AI Video Prompt?

By MakeThisVid Editorial Updated Methodology

This one surprises people because the name is almost always innocuous — a common first name, used once, in a line of dialogue between two fictional characters. The clip fails anyway. Rerun the exact same prompt and it might succeed. That inconsistency is what throws people off: they conclude the tool is broken, or that their scene was somehow inappropriate, when the actual cause is a single word sitting in the prompt the whole time. Video models run likeness and real-person protections on every request. A name is treated as a possible reference to a real, identifiable person, and the render can be declined on that basis — regardless of whether the person in your photo is that person, and regardless of what you meant by it.

Make a video from one prompt. $2.99 a video · Ready in about a minute · No subscription.

Key facts

Trigger Any name Common first names count — the person doesn't need to be famous
Highest-risk spot Dialogue vocatives "Hey Matt," is riskier than a plain "Hey,"
Behavior Nondeterministic Same prompt can pass on one run and fail on the next
Fix Describe by appearance "the woman in the red coat" instead of a name
Cost of the fix Usually one word Cutting a vocative from dialogue rarely changes the line

How to Use MakeThisVid

From prompt to downloadable MP4, ready to deploy.

  1. Understand what's actually happening

    The video model checks every prompt for possible references to real, identifiable people. A name is enough to count as a possible reference, so the request can be declined on that basis alone — it doesn't matter whether the person is famous, whether your uploaded photo is of them, or what you intended the scene to mean.

  2. Identify characters by appearance and role, never by name

    Swap the name for what the character looks like or does: "the woman in the red coat," "the older man," "the barista." This isn't a workaround, it's a better instruction either way — it also tells the model what to actually render, where a name tells it nothing about appearance.

  3. Cut the vocative in dialogue

    Most failures trace back to a name inside a line of spoken dialogue, not the scene description. "Hey, are you coming?" works where "Hey Matt, are you coming?" may not. Dropping the vocative is usually a one-word deletion and costs the line nothing.

  4. If a clip fails for no obvious reason, remove any name first

    Before you touch the camera angle, lighting, or motion, delete every proper name in the prompt and rerun. It's the cheapest single change to test. Because the trigger is nondeterministic, a plain rerun with no changes can also happen to succeed — removing the name first is how you tell the difference between bad luck and a real cause.

Make a video — $2.99

Who Uses MakeThisVid for This

Multi-character dialogue scenes

Any prompt where one character addresses another by name in a line of dialogue is the highest-risk pattern — it's the one most people write without thinking twice, because it reads naturally on the page.

Branded or scripted ad reads

A script that names a spokesperson, a customer, or a brand character carries the same risk as any other dialogue. Appearance-based descriptions carry the same information without the trigger.

Photo-to-video with a named subject

Naming the person in an uploaded photo doesn't make the model more accurate about who they are — it just adds a name-shaped trigger to the prompt. Describe them by what's visible in the photo instead.

Frequently Asked Questions

Names are checked as possible references to real, identifiable people. When the model can't rule that out, it can decline the render — even for an ordinary first name in a line of fictional dialogue, and even when the photo you uploaded isn't of anyone by that name.
No. Common, everyday first names are enough to trigger it. The protection isn't specifically about celebrities or public figures — it's about any name that could plausibly belong to a real, identifiable person, which is most names.
The check is nondeterministic. The identical prompt, containing the identical name, can pass on one run and get declined on the next. That inconsistency is exactly why the pattern is easy to miss — a single clean rerun can look like proof the name wasn't the problem, when it was.
No. This isn't about the content of the scene — it's specifically about whether a name could refer to a real person. A perfectly ordinary, wholesome scene can fail for exactly this reason, and removing the name typically resolves it without changing anything else.
Describe the character by appearance or role: "the woman in the yellow scarf," "the older man," "the barista." This also gives the model more to work with visually, since a name carries no information about what the character looks like.
No. Removing the name removes one known trigger, not every possible reason a render can fail. If a clip still fails with no name anywhere in the prompt, the cause is something else — but a name is the first and cheapest thing to rule out.

Write your next prompt without the trigger

$2.99 makes one clean video, ready in about a minute — or add 3 more takes for $6.99 — any photo, any look, same clean 1080p with commercial use. Pay once. No subscription required.

Make a video — $2.99

Broken renders are remade or refunded.

or 3 more takes — $6.99 Any photo, any look · About $2.33 a take · No subscription required.