Polaris
BlogAnswer

Can You Use AI Actors in Ads? Yes, With Some Guardrails

A clear answer on whether AI actors are allowed in ads, plus the practical guardrails around scripts, disclosure, and claims.

AnswerJul 15, 2026·2 min read

Yes, you can use AI actors in ads, and it's an increasingly common way to produce UGC-style video without booking a shoot. The key distinction that matters is what the AI actor is doing: delivering a script you wrote, versus impersonating a specific real, named person without their consent. The former is standard practice; the latter runs into both platform policy and, depending on jurisdiction, legal issues around likeness rights.

What's actually allowed with AI actors?

Generating an AI talking-actor delivering your own ad script, built around your own product, is the normal use case advertisers reach for. Tools that generate this kind of content, including Polaris's Omni talking-actor model, work from your script and product, not from a library of cloned real people. That distinction is worth checking whenever you evaluate a tool: does it build around your product and your words, or does it lean on cloning identifiable real individuals? The former is the safer, more standard lane.

Do platforms allow AI actors in ads?

Generally yes, but most major platforms now require disclosure when the AI content is realistic enough that a viewer could mistake it for unaltered real footage. That's a labeling requirement, not a ban on using AI actors - check each platform's current policy (TikTok, Meta, etc.) before launching, since rules and enforcement shift over time.

What should the actor actually say?

Whatever your product genuinely supports. Regulated categories especially need care here - avoid disease or cure claims, guaranteed-result claims, or anything the product can't back up, regardless of whether a human or AI delivers the line. An AI actor doesn't loosen compliance standards; if anything, it's worth double-checking the script since it's easy to generate several variations quickly and accidentally let an overreaching claim slip through.

How do advertisers actually put this into practice?

Write (or paste) the script, attach a product photo, and let the generation handle the delivery - with Polaris that happens directly in a Claude chat, with a live panel rendering the finished ad in place. From there it's a normal testing loop: generate a few script and hook variations, track hook rate and ROAS with the hook generator and ROAS calculator, and keep the winners running. See generating ads inside Claude via MCP and the AI UGC ads overview for more on the workflow.

Frequently asked questions

Is it legal to use AI actors in ads?
Generally yes when the AI actor is delivering your own script and isn't cloning a specific real, identifiable person without consent. Likeness and platform rules vary, so verify for your jurisdiction and platform.
Do I need to disclose that an ad uses an AI actor?
Most major platforms require disclosure when AI content is realistic enough to be mistaken for real footage. Check each platform's current AI content policy before publishing.
Can an AI actor make any claim I want in an ad?
No. The same advertising compliance rules apply regardless of who or what delivers the line - avoid disease, cure, or guaranteed-result claims, especially in regulated niches.
How is an AI actor different from a cloned real person?
An AI actor delivers a script you provide, built for your product. Cloning a specific named real person's likeness without consent is a different, more legally sensitive practice.

Create your first winner now

One product link in. Winning ads out.

Keep reading

AnswerSound-On vs Sound-Off Ads: How to Design for BothAnswerSquare vs Vertical Video Ads: Which Format to UseAnswerDo Ads Need a Person in Them? What Actually Matters