Is an AI influencer legal? Platform rules, disclosure and likeness

Anyone asking whether an AI influencer is legal will find rules of three kinds: the platforms' (AI labels), advertising's (marking paid posts as ads) and those on other people's faces, voices and footage. This page reports what each source states, with its link and date, and draws no legal conclusions. It is not legal advice.

What the rules cover

The advertising and election rules below are Brazil's, the one country whose written rules we read. The platform rules are the platforms' own pages. Anyone selling ads in another country should check that country's advertising regulator.

Disclosure: what YouTube, Instagram and TikTok ask

Each platform words its rule its own way. All three ask for a label when AI content looks real.

YouTube

YouTube asks you to disclose when AI meaningfully alters or generates photorealistic content: a real person made to say or do something they never did, footage of a real event or place altered, a realistic scene that never happened. According to the page, non-realistic or animated content, fantastical scenes, beauty filters, upscaling, AI-written scripts and thumbnails, and cloning your own voice for narration do not need a label. The label goes in the player for photorealistic content and in the expanded description for the rest. Skip it repeatedly and YouTube may label the video itself or penalize you, up to removing content or suspending you from the YouTube Partner Program. The page also says disclosing does not limit a video's audience or its eligibility to earn money (YouTube help page, read October 2026).

Earning on YouTube falls under a separate policy. On July 15, 2025, "repetitious content" was renamed "inauthentic content". The page lists "using AI to visualize a unique character and narrative you invented" among the allowed uses, and says channels with AI personas that present themselves as human experts giving advice on sensitive topics, such as an AI "doctor" or legal advisor, will not be allowed to monetize (YouTube policy). What that means for an account made of trends is in faceless TikTok with an AI character.

Instagram and Facebook

Instagram and Facebook fall under Meta's Community Standards. Meta asks for disclosure through its AI-disclosure tool on organic posts with photorealistic video or realistic-sounding audio, says it may penalize people who do not disclose, and may add a label to high-risk content (Meta Community Standards). The label reads "AI info" (it was "Made with AI" until July 1, 2024) and appears either when Meta detects signals the industry shares or when the person discloses (Meta, April 2024, updated through October 2025).

On August 31, 2026, Instagram announced that its "AI creator" label is renamed "AI-generated profile". Creators add it by editing the profile, and Instagram says it will limit the reach of profiles that should carry it and do not. The guide to AI labels on Instagram, TikTok and YouTube has the details.

TikTok

TikTok said in May 2024 that it labels AI content uploaded from some other platforms automatically, using C2PA Content Credentials, that creators must label realistic AI-generated content, and that anything made with TikTok's own effects is labeled for them. Its policies also prohibit harmfully misleading AI content whether or not it is labeled (TikTok Newsroom, May 9, 2024).

Platform pages change. Read each one before you post.

Paid posts: what advertising rules ask

The one written advertising rule we read in full is Brazil's. CONAR, the country's advertising self-regulation council, updated its guide to advertising by digital influencers in 2026, and the guide's definition of influencer advertising covers third-party content that promotes products or brands on the profile of a natural person, a virtual persona, animals, avatars or other characters or themes, computer-generated ones included, under reciprocal commitments between advertiser and influencer (CONAR guide, 2026 edition).

Section 1.3.1 is about AI. The CONAR code (the CBAP) and the guide apply in full to advertising whatever technology produced and published it, AI-generated, edited or targeted content included. Influencers, advertisers and agencies stay responsible for the result, and an ad may not contain simulations, endorsements or testimonials that could mislead consumers about the offer. A footnote says the guide adds no new duty to disclose that AI was used, and that CONAR follows international rules and is reviewing cases in its Ethics Council (same guide).

To mark an ad, the guide wants paid content clearly identified as advertising, preferably with the platform's own paid-content feature, or with terms such as "#publicidade" or "#publi", visible up front, with no tap on "more" needed. The guide on brand deals with an AI influencer shows how that fits into a deal.

A second footnote says a virtual influencer can be a fictional computer-generated "person" or avatar with realistic human traits, and that international guidance points to telling consumers they are not interacting with a real human. For children and teenagers it cites article 11 of Decree 12.880/26, which calls for transparency about the synthetic, automated nature of an interaction, in the cases the article lists. CONAR's guides page shows only this influencer guide and one on children's online advertising, with no separate AI guide (conar.org.br/diretrizes, read October 13, 2026).

One outside view, from before the 2026 guide: in June 2025 Meio & Mensagem spoke to Ricardo Cavallini of Singularity Brasil. According to the report, he noted that Brazil had no case law yet on the use of AI in advertising work, and said that "embora ainda não exista uma lei específica, o ideal é que o uso de IA seja claramente sinalizado" (although there is no specific law yet, the ideal is for AI use to be clearly flagged). He closed with "A transparência é gratuita e evita qualquer ruído futuro" (transparency costs nothing and avoids future noise) (Meio & Mensagem, June 12, 2025).

Faces, voices and footage that belong to other people

We could not open the text of Brazil's federal statutes on image rights. This section reports what rules and news reports say, and what the law requires in a given case is left out.

For testimonials and the use of people's image in advertising, CONAR's guide points to articles 27 (paragraph 9) and 34 and Annex Q of the CBAP (footnote 7 of the guide).

On footage, the question reached the press in October 2026. The Brazilian paper O Tempo reported on October 8 about the wave of Jean Phil videos (who he is) and about a Brazilian account that, according to the paper, uses AI to "appear" as Vector from Despicable Me in videos of well-known creators and artists (how those videos work). The report says the practice raises questions of copyright, image rights and commercial exploitation, and writes that the recording, performance, dance or editing that produced the scene still belong to third parties even when AI swaps the appearance. It quotes no lawyer (O Tempo, October 8, 2026). No court decision on these cases is in the sources we read.

Elections

Election propaganda has its own rules. Jornal do Brasil reported in August 2026 that Resolution 23.610/2019 of Brazil's electoral court (the TSE) bans electoral use of synthetic audio or video that creates or manipulates the image or voice of living, dead or fictitious people, that permitted AI uses must be disclosed explicitly and prominently, and that a label does not excuse a prohibited use. In the same report TikTok said creators must label synthetic content and that it removes violating material even when labeled (Jornal do Brasil, August 29, 2026). BBC News Brasil, in April 2026, quoted Professor Yasmin Curzi of FGV Direito Rio as saying that the new TSE resolution (23.755/2026) broadened and tightened the rules on deepfakes. The same report says the TSE approved this year a measure banning AI in electoral propaganda in the 72 hours before the election and the 24 hours after, without saying whether it is part of that resolution (BBC News Brasil, via O Povo, April 21, 2026). The two reports cite resolutions with different numbers (23.610/2019 and 23.755/2026), and this page did not read the text of either.

A political-commentary AI account, Dona Maria, has already reached that court. On April 22, 2026, three parties (PT, PV and PCdoB) asked the TSE to suspend all her profiles, obtain the information to identify those responsible for them and for their monetization, and treat the content as illegal, citing among other points the use of AI without identification (Terra, April 27, 2026). The outcome is not in the sources we read.

Where Brazil's AI bills stand

None of them is law. Bill 2338/2023, the AI framework, sits in the Chamber of Deputies. The Chamber's tracking page showed on October 13, 2026 the status "Aguardando Parecer do(a) Relator(a) na Comissão Especial", which means the special committee is waiting for its rapporteur's opinion. The committee was created on April 4, 2025 and constituted on April 29, Deputy Aguinaldo Ribeiro (PP-PB) was named rapporteur on May 20, 2025, and the page shows no approval by the Chamber. Thirty-seven other bills are attached to it, the latest movement dated September 1 and 2, 2026 (Chamber of Deputies).

Some of the attached bills, all proposals, by subject:

The page was truncated when we read it, so the list may be longer.

What Duble AI asks of you

Duble AI makes videos with an AI influencer in place of someone, so the three questions above come up for the people who use it. This is what the app says and asks.

• Labels: a finished video's page reminds you to mark it as AI-made content when you post, and the public page of a shared video says "Video made with AI". Videos come out with no watermark, so the label on the platform is up to you.
• People: the Terms do not allow using another person's name, image or appearance without their permission, and never those of celebrities or political figures. When you make an influencer from photos, you tick a box saying the images are yours or that you have permission to use them. The Terms ask that an invented influencer resemble nobody real, and say you are responsible for what you upload, generate and post.
• Videos: when you send a video of your own, the app reminds you to send only videos you have the right to use. The trends in the library are other people's videos as they were posted, and whoever posts is responsible for what they post.
• Politics: the check on what you type refuses politicians, parties, candidates, elections and campaigns, even as an influencer's name.

The rule about celebrities and other people's image is in the Terms, and keeping it is up to the user. To make an invented influencer from scratch, create yours. The guide to an AI influencer from your own photo covers which photos to send and the rights involved.

What this page leaves out

Copyright in the music, choreography and videos that appear in trends, laws outside Brazil, brand contracts and the text of the federal image-rights statutes are not covered. For any of them, the next step is a lawyer.

This is not legal advice

This page gathers what the cited sources said in October 2026, with the date of each. Platform rules change and bills can move at any time. For your own case, talk to a lawyer. This is not legal advice.