AI Video & Commercials

Disclosing AI in Advertising: Ethics, Platform Rules and Trust

Do you have to tell audiences a video was made with AI? A guide to platform labelling rules, Australian Consumer Law risk and sensible disclosure habits.

There is no single law that says every AI-assisted video needs a label. But two things do apply: platforms set their own disclosure rules for realistic AI content, and Australian Consumer Law (ACL) prohibits advertising that misleads. Whether you must disclose depends on what the AI-made content shows and what viewers would reasonably assume. This is general information, not legal advice.

Three separate questions

People often mix up three different obligations. Treat them separately.

  1. Platform rules. Does the channel require you to flag the content when you upload?
  2. Consumer law. Does the ad, taken as a whole, create a false impression?
  3. Trust. Would your audience feel misled if they found out later?

An ad can pass the first and still fail the second or third.

What platforms ask for

YouTube’s help page says creators must disclose AI-generated or altered content that seems realistic. Its examples include showing a real person saying or doing something they did not, changing footage of a real event or place, and creating a believable scene that never happened. Non-realistic content and minor edits, such as beauty filters, captions, upscaling or audio repair, do not need disclosure (YouTube Help).

Labels differ by content type. YouTube places a label in the video player for photorealistic AI content or AI-generated music, and in the expanded description for other cases. It can also apply labels automatically, including where content carries C2PA metadata (YouTube Help).

Meta describes an “AI info” label that it adds when it detects industry-standard AI indicators or when people disclose AI use. Its page says the approach is still evolving and may miss some AI-edited content (Meta Transparency Centre).

Platform rules change often. At the time of writing (November 2026) these pages say what is quoted here, so check the current version before each campaign.

The consumer law angle

The ACCC explains that conduct can mislead even when the business did not intend it to. The test is the overall impression a statement creates, and leaving out important information can also mislead. Claims should be accurate and backed by evidence (ACCC).

AI use is not itself the problem. The risk comes from what the video implies. Some examples of where the impression could be false:

  • A synthetic “customer” presented as a real person giving a real testimonial.
  • A product shown performing in a way the real product cannot.
  • A rendered location or result presented as documentary footage.

An AI-generated mascot or a clearly stylised scene is different: no reasonable viewer would assume it is a recording.

A practical disclosure decision table

Situation Suggested approach
Clearly stylised or animated, no real people Platform label only if the platform asks; no extra notice usually needed
Photorealistic scene that never happened Use the platform label; consider a short on-screen or caption note
Real person’s likeness or voice recreated Get written consent, label it, and keep the consent on file
Testimonial or review content Use real, attributable statements only; never synthesise a customer
Product demonstration Show the real product behaviour; mark concept visuals as such

Habits that build trust

  • Keep a one-page record per video: tools used, what was generated, what was filmed, who approved it.
  • Write the caption note in plain words, for example “Scenes in this ad were created with AI.”
  • Do not strip content credentials from files you publish.
  • Decide your disclosure policy before production, so it is not a last-minute argument.

Limits of this guidance

Platform wording, ACCC guidance and any future Australian AI rules can change. If a campaign involves real people’s likenesses, health or financial claims, or comparisons with competitors, get advice from a lawyer who handles advertising law.

For the wider picture of where AI video helps and where it does not, see the real limitations of AI-generated video in commercial production, and for a broader introduction read AI video production for brands. If product visuals are involved, AI product videos covers realistic expectations.

Where Radiance Impact fits

Radiance Impact’s Studio produces AI commercials, product films and short-form video, and can help you plan what is generated, what is filmed and how it is labelled. To talk through a project, contact us.

Sources