Skip to content
Entertainment

Unauthorised use of an artist's image in Brazil: what compensation is worth and who pays for the fake ad

The brand that keeps running a campaign after the contract ends, the company that reposts an influencer's photo, the deepfake ad with a celebrity's face: what an artist can claim under Brazilian law, how courts calculate material damages, the disgorgement of profits made with the image, platform liability for paid ads and the deadline to act.

By Cesar Augusto Taborda Lima

Whoever uses an artist’s image to sell, without authorisation, must compensate the artist, and the artist need not prove loss to recover. The claim can have three parts: moral damages; material damages, which courts measure by the fee the artist would have charged; and, where the company profited from the use, the return of that profit. The most frequent cases today are the campaign that stays online after the contract ends and the fake ad with a celebrity’s face, in which the platform now answers on a presumption of fault when the ad is paid. The deadline to claim is, as a rule, three years.

Compensation for the unauthorised publication of a person’s image for economic or commercial purposes does not depend on proof of loss.

Superior Court of Justice, Súmula 403, Second Section, 2009 (free translation).

The campaign that never comes down

A model was hired by a Belo Horizonte clothing brand for a six-month campaign, for R$18,000. The contract ended in August 2023. The brand and a football club, partners in a collection, kept posting his photos on the club’s profile for about seven months. In November 2025 the Minas Gerais Court of Appeal ordered the companies to pay R$18,000 in moral damages and R$21,000 in material damages, a figure equal to R$3,000 for each month of unauthorised use, exactly the monthly price of the expired contract (Civil Appeal 1.0000.25.145018-5/001).

The calculation follows a line the Superior Court of Justice had already drawn. In 2015 its Third Panel set material damages at what the claimant would have received had the contract been renewed, reduced in proportion to the actual use, and explained why: failing to make good what he did not earn would amount to unjust enrichment of the company (REsp 1,323,586/PB, Justice Ricardo Villas Bôas Cueva reporting). The old contract ends up serving as the price list for the unlawful use.

The same logic applies to the influencer whose photo appears on a shop’s commercial profile, the singer whose video becomes the ad for an event that never hired him and the actress whose soap opera still illustrates a clinic’s ad. A repost with credit is still commercial use without authorisation, and it is the commercial purpose that brings in Súmula 403.

What it is worth

Moral damages in these cases are presumed. The amounts courts award tend to be moderate, and they weigh the artist’s profile, the duration and reach of the use and any association with a product foreign to the artist’s work or contrary to their image. In the Minas Gerais case, it counted against the companies that the campaign fell outside the model’s market segment. A singer known for a healthy-living message, tied to a miracle weight-loss pill, suffers greater harm than a neutral photo in a shop window would cause.

Material damages require proof, and the best proof is the artist’s own market: earlier contracts, offers received, the agency’s rate card, the media kit. An artist who cannot show what they charge leaves the judge to guess, and judges tend to guess low.

The third part is the least known and can be the largest. In 2018, in the case of an actress whose image a pharmacy used in advertising, the Superior Court of Justice recognised the disgorgement of profits from intervention: whoever exploits another’s image without authorisation must return the profit made from it, on the basis of unjust enrichment (REsp 1,698,701/RJ, Justice Ricardo Villas Bôas Cueva reporting). The amount is set by an expert at the liquidation stage, by the company’s actual profit during the period of use, less what is owed to other factors, such as the advertiser’s own structure and work. The claim needs at least some indication that there was a gain. In the Minas Gerais case it was rejected precisely because the claimant did not show, even minimally, that the companies had been enriched. Campaign sales figures, orders placed with the ad’s discount code and the paid reach of the post are the start of that proof, and what lies in the company’s hands can be ordered produced in the proceedings.

The fake ad with a famous face

The fastest-growing case is a different one. A sponsored video shows the face and voice of a well-known artist, altered by artificial intelligence, recommending an unregistered remedy, a course or an investment. The advertiser is liable for everything, but is usually a nameless profile, sometimes run from abroad. Identifying it depends on a court order requiring the platform to hand over its records, under art. 22 of Brazil’s Internet Civil Framework, as I explained in the article on fake profiles and online scams.

The platform has become a viable defendant. In reinterpreting art. 19 of the Internet Civil Framework, Brazil’s Supreme Court held that a platform is, as a rule, jointly liable for unlawful third-party content it keeps online after notice and that, for paid ads and boosted posts, its fault is presumed even without notice, unless it proves it acted diligently and in reasonable time (Themes 987 and 533, final thesis of 17 June 2026). The rule applies from 5 August 2025 and reaches content that remains available.

The date matters. In September 2026 the Federal District Court of Appeal decided the case of an endocrinologist whose face and voice were used in a sponsored Instagram video to sell a supplement banned by Brazil’s health regulator. She obtained removal and the profile’s data, but compensation from the platform was denied, because the facts predated the Supreme Court’s change of position and the video came down after the court order (case 0714186-75.2024.8.07.0001). The same ad, run today, would reach the judge with the platform’s fault presumed. Meta itself announced in February 2026 lawsuits against advertisers who used celebrities’ faces in scams in Brazil, which shows the problem is known to those who sell the space.

For the artist, the central proof is that it was a paid ad. The platform’s public ad library shows the advertising page and the date the ad started running, and a capture of it, taken with date and address, is worth more than a capture of the video alone. If the deepfake puts words offensive to the artist’s honour in their mouth, that specific part follows the old art. 19 regime, which requires a court order to make the platform liable; the commercial use of the face remains under the general rule. I covered voice and image clones, and the contract clauses they call for, in the article on artists’ voices and images recreated by AI.

What does not give rise to compensation

A photo of an artist at a public event, published by a news outlet to inform, needs no authorisation. In deciding ADI 4,815 in 2015, the Supreme Court also removed the need for authorisation for biographies. The position changes when the same news photo is reused in an ad, because the purpose then becomes commercial.

There is also the difference between the photograph and the face. The photograph is the photographer’s work, protected by Law 9,610/1998 (art. 7, VII), and the image portrayed belongs to the person. A brand that buys a licence for a photo from an image bank acquires the photographer’s rights and still needs the artist’s authorisation to use their face in advertising; the law itself restricts the photographer on this point (art. 79). The reverse also holds: an artist who uses a photo taken by someone else in their own campaign needs a licence from whoever took it.

How to act

The prudent order starts with evidence: captures with address and date, the ad library page where there is one, and a notarial record if the case is significant, because ads change and disappear. Next, notice to the brand, which in many expired-campaign cases ends in a retroactive licence agreement, cheaper for both sides than litigation. Notice to the platform must give the exact address of each piece of content, and its record is what marks, from then on, the liability of whoever keeps it online. If none of that works, the lawsuit seeks urgent removal, compensation and, where needed, the advertiser’s data. The claim for compensation is time-barred, as a rule, after three years (Civil Code, art. 206, § 3, V), and platforms need only keep the records that identify an advertiser for six months.

It seems to me that the most effective protection is the one the artist writes in advance. An image contract that sets term, territory, media and a price per month of use beyond what was agreed turns the campaign that never comes down into a simple calculation, at a price the brand has already accepted. And proof of one’s own fee, kept with care, is what separates a token award from one that pays what the image is worth.