The artificial intelligence generating these videos has crossed a threshold where the human eye cannot reliably distinguish the real from the fabricated.
In Rio de Janeiro, beloved television presenter Xuxa Meneghel has turned to the courts to reclaim something that should never have required reclaiming: her own face. The lawsuit she filed against Bagy Digital Commerce Solutions — seeking 100,000 reais after her likeness was fabricated into an AI-generated advertisement — is less a story about one celebrity's grievance than about a civilizational lag, the moment when technology outran the legal and perceptual tools humanity built to protect identity and truth. At a time when four out of five Brazilians have already encountered deepfakes yet fewer than one in three can detect them, the case asks a question that will only grow more urgent: who owns a person's image when machines can conjure it without permission, and without a trace?
- A company deployed a fabricated video of Xuxa endorsing its AI sales tool, counting on her trusted face to deceive consumers who had no reason to doubt what they were seeing.
- Brazil is among the world's most deepfake-exposed nations, with 80% of its population having encountered manipulated media — yet only 29% can reliably tell the real from the fabricated.
- The visual errors that once made deepfakes detectable — distorted hands, flickering features, unnatural motion — have largely vanished, leaving ordinary viewers without the tools to protect themselves.
- Xuxa's lawsuit represents one of the few available responses: a legal claim for damages that forces a company to answer for the theft and weaponization of a person's likeness.
- The case lands in a gap between what the technology can do and what the law, detection methods, and regulatory frameworks are currently equipped to prevent or remedy.
Xuxa Meneghel, one of Brazil's most recognized television personalities, has filed suit in Rio de Janeiro state court seeking 100,000 reais in damages after her image was used without consent in a deepfake advertisement. The company, Bagy Digital Commerce Solutions, inserted a manipulated version of the presenter into a promotional video for an AI sales tool — exploiting her reputation to lend false credibility to a product she never agreed to endorse.
The case arrives against a backdrop that makes it feel less like an isolated incident and more like a warning. Research by digital identity firm Veriff found that 80 percent of Brazilians have already encountered deepfakes online — well above the global average — yet only 29 percent could correctly identify a manipulated video when shown one. The technology has crossed a threshold: the crude visual errors that once gave deepfakes away have largely disappeared, and the human eye, untrained, can no longer be trusted to tell the difference.
Veriff's director of emerging markets put it plainly: two years ago, people believed they could spot a fake because the mistakes were obvious. Today those mistakes are gone. Participants in the study were asked to sort authentic videos from AI-generated imagery, manipulated photos, and face-swapped deepfakes — and the results revealed how exposed ordinary people have become.
Xuxa's lawsuit is a legal attempt to hold one company accountable, but it also illuminates a structural problem. Detection methods trail creation methods. Laws have not kept pace with the technology. And public figures — whose faces are widely available and whose reputations carry real commercial value — have limited recourse once their image has been taken and put to use without their knowledge.
The 100,000-real claim is modest against the true cost of losing control over one's own likeness. But it marks a threshold of its own: the moment a prominent entertainer had to go to court to reclaim ownership of her face. As deepfake technology grows more capable and more accessible, it is unlikely to be the last such moment.
Xuxa Meneghel, one of Brazil's most recognizable television personalities, has filed a lawsuit in Rio de Janeiro state court seeking 100,000 reais in damages after discovering her image had been used without permission in a deepfake advertisement. The company behind the scheme, Bagy Digital Commerce Solutions, deployed a manipulated video of the presenter to promote an artificial intelligence sales tool, banking on her face and reputation to lend false credibility to the pitch. The deception worked by design: viewers seeing what appeared to be Xuxa endorsing the product would naturally assume she had authorized the promotion.
The case arrives at a moment when deepfake technology has outpaced both human perception and legal protections. A 2025 study by Veriff, a digital identity verification firm, found that 80 percent of Brazilians have already encountered deepfakes on social media and online platforms—well above the global average of 60 percent. Yet the same research revealed a troubling gap between exposure and detection: only 29 percent of Brazilian respondents could correctly identify a manipulated video when shown one, and just 35 percent reliably recognized authentic content. The technology has become so sophisticated that the visual tells have largely vanished. Where deepfakes once betrayed themselves through obvious glitches—extra fingers, warped facial features, unnatural movements—those markers have nearly disappeared.
Andrea Rozenberg, Veriff's director of emerging markets, described the shift bluntly. Two years ago, she noted, people believed they could spot a fake because the errors were visible and crude. Today those errors are gone. The artificial intelligence generating these videos has crossed a threshold where the human eye, untrained and unprepared, cannot reliably distinguish the real from the fabricated. The research tested participants by asking them to sort through a mix of authentic videos, AI-generated imagery, manipulated photos, and deepfakes created by swapping faces into real footage. The results showed how vulnerable ordinary people have become.
Xuxa's lawsuit is one response to this vulnerability—a legal attempt to hold a company accountable for theft of likeness and the deliberate deception of consumers. But the case also exposes a deeper problem. The technology that created her deepfake moved faster than the laws designed to protect her. Detection methods lag behind creation methods. Regulatory frameworks have not caught up. And the people most at risk—public figures whose faces and voices are widely available, whose reputations carry commercial weight—have limited recourse once their image has been stolen and weaponized.
The 100,000-real claim is modest compared to what such a violation might cost Xuxa in lost control over her own image and the erosion of trust between her and her audience. But it is also a marker: the moment when one of Brazil's most prominent entertainers had to go to court to reclaim ownership of her own face. As deepfake technology continues to improve and spread, more cases like this one will likely follow.
Notable Quotes
Two years ago, people believed they could spot a fake because the errors were visible and crude. Today those errors are gone.— Andrea Rozenberg, director of emerging markets at Veriff
The Hearth Conversation Another angle on the story
Why does it matter that Brazilians are more exposed to deepfakes than the global average?
Because Brazil is a test case. If 80 percent of people there have seen deepfakes, that's not a fringe problem—it's the new normal. And if only 29 percent can identify them, you're looking at a population where most people are regularly fooled.
But couldn't Xuxa just post a denial and move on?
She could, but that doesn't undo the damage. Someone saw that fake video and believed she endorsed a product. Her reputation was borrowed without permission. A denial reaches far fewer people than the original lie.
The study mentions that deepfakes used to have visible errors. What changed?
The AI got better. The models trained on millions of faces, the rendering improved, the processing power increased. What took hours to create now takes minutes. The tells—the glitches that used to give it away—have been engineered out.
Is a lawsuit the right tool for this problem?
It's one tool. It establishes that there's a cost to stealing someone's image. But the real problem is that detection hasn't kept pace with creation. By the time Xuxa's lawyers proved the deepfake was fake, thousands of people had already seen it and believed it.
What happens to the next person this happens to?
They'll probably sue too. But until there's a way to watermark authentic content or detect deepfakes reliably, the technology will keep outrunning the law.