Last week, the Brazilian internet met talk show host Marisa Maiô—a plump woman wearing nothing but a black one-piece swimsuit and high heels, who hosts a typical afternoon talk show seen on broadcast television. With biting humor and blunt commentary, Marisa presents a set of highly questionable attractions: from elderly women competing to be the first duped in a phone scam, to a mother forced to hand over her child as punishment for failing a quiz question—among other absurd situations. As it turns out, Marisa, her show, and her guests are all creations of artificial intelligence, produced with Google’s new video generator, Veo 3, based on an idea by screenwriter Raony Phillips.
The video, edited to resemble a compilation from a real TV program, quickly went viral across digital platforms and messaging apps like Telegram and WhatsApp, where it was marked as “forwarded many times,” indicating massive reach. Its technical merit is undeniable; the facial expressions, colloquial Brazilian Portuguese, accent, and prosody are so convincing that even those familiar with synthetic videos might overlook its artificial nature.
The week before, similarly realistic videos were already circulating in WhatsApp groups, also with high virality. These featured biblical characters as digital influencers—Noah complaining about the challenges of building the ark, a Hebrew narrating the Red Sea’s parting live, and the Virgin Mary’s pregnancy saga en route to Bethlehem. The difference in those cases was that the historical impossibility of mobile phones in ancient times made their fictional nature evident.
What I want to highlight is not only the high plausibility of these videos and their growing indistinguishability from reality with each new generation of AI. Generative programming videos have fueled intense debate about the future of audiovisual media. Aesthetically, they open up new creative possibilities, allowing the production of complex effects at minimal cost thanks to tools that were once exclusive to major studios. This democratization enables linguistic and narrative experimentation that was previously unimaginable.
Legal and labor challenges
On the legal front, pressing questions arise around copyright, as seen in the case of Marisa Maiô, whose character was replicated by third parties in new videos—and even used by a retail chain in advertising campaigns—without compensation to the original creator.
There is also increasing concern about the impact on the job market. The entire audiovisual production chain, from actors to technical crews, editors, and writers, faces potential displacement due to creative automation.
Virtual sensationalism
But what happens when these videos cross into territory where the boundaries are no longer technical but fundamentally ethical? Popular talk shows have a history of transgressing such boundaries, deliberately creating a gray zone between entertainment and information with sensationalist skits, exaggerated emotional appeals, and frequent trivialization of violence—verbal or physical. As analyzed by Muniz Sodré in The Communication of the Grotesque, the bizarre in media acts as a symbolic language that destabilizes boundaries between the real and the fictional, the ethical and the sensationalist. These shows create a space of controlled chaos, where social dysfunction becomes spectacle.
Still, traditional radio and TV, despite their frequent ethical excesses, have always faced some scrutiny when absurdity and grotesqueness crossed the line. Think back to the fierce Sunday ratings battle in the 1990s between Fausto Silva (Globo) and Gugu Liberato (SBT), which led to tasteless spectacles, or the older rivalry between Chacrinha and Flávio Cavalcanti in the 1970s.
However, such abuses were somewhat curbed by the fact that TV and radio licenses are public and regulated. This mechanism, though imperfect, enforced a degree of self-control by broadcasters, driven by the need to preserve their public image, appease advertisers and regulators, and maintain some degree of institutional credibility.
An algorithmic update of grotesque TV
The fictional show Marisa Maiô precisely satirizes the thin ethical lines media companies are willing to cross in their pursuit of viewership. In this sense, Marisa Maiô updates the tradition of grotesque television in algorithmic terms. The use of scandal as a tool to maintain public attention, described by Danilo Angrimani in Press Hard Enough and Blood Will Flow (1994), is part of a “drama of pain” structured around the logic of sensationalism. The difference with synthetic videos is that this logic is automated by systems whose sole metric of success is engagement—freeing the content entirely from reality or ethical accountability.
In Marisa Maiô’s video, the ethical provocation peaks when two guests break into an all-out brawl, while the host cynically remarks, “What’s great here is that there’s no security and we won’t stop them.” The spectacle of absurdity reaches new heights in another scene, where a grieving woman stands beside a coffin covered in black cloth. Miriam, another host, announces that a recently deceased relative lies within—and that their identity will be revealed live to the audience.
Normalization through trivialization of the absurd
Beyond the blurred line between simulation and reality, the episode reveals a deeper concern: how frequent and convincing depictions of absurdity may normalize it socially, making it acceptable. With no technical or ethical regulation, no concern for corporate credibility, and no fear of commercial backlash, algorithmic production can cross moral boundaries without hesitation—trivializing grotesqueness and cynicism as social norms.
In fact, today’s digital platform business model favors extreme content, as monetization depends on engagement and audience attention—without any ethical consideration for content quality. This economic incentive all but guarantees the rise of even more extreme videos grounded in absurdity—and not necessarily humor.
Digital networks’ hyper-segmentation may even allow customized content tailored to the audience’s darkest desires. Police news segments, which already expose violence and anti-human-rights rhetoric, could go further, showing artificially created footage of executions and torture of cybercriminals.
Similarly, the vast sexual range of the internet could harbor disturbing material, including extreme sexual violence and pedophilia, in synthetic formats. Fake news programs could present fabricated stories about political figures or socially marginalized groups, exploiting pre-existing confirmation biases—all under the guise of “just a simulation.” While there is growing advocacy for labeling AI-generated videos to alert viewers they are not real, this is unlikely to curb the cultural trend of embracing disinformation.
The perversity of knowing no one was “really” harmed
There’s an inherent perversity in knowing no real humans were involved—so no one was actually tortured, raped, or publicly shamed for false accusations. That knowledge becomes a convenient excuse for consumption and sharing, free of guilt or revulsion. No matter how violent the video, “no one is really suffering.”
While deepfakes have already sparked concern over truth erosion, the Marisa Maiô case suggests something deeper: the trivialization of ethical deviation as something funny, entertaining, or even welcome. With increasingly accessible tools, anyone can now create hyper-realistic videos that normalize the unacceptable.
A culture molded by absurdity
The risk goes beyond spreading disinformation—it also involves shaping a subjectivity based on absurdity, where violence, cruelty, and senselessness become acceptable parts of daily life and visual culture. This, inevitably, has practical consequences for citizenship.
As these tools proliferate unchecked by ethical constraints, it is foreseeable that the internet will become flooded with content that defies ethical limits, contributing to the normalization of looser moral standards. The consequences of a subjectivity shaped by absurdity are unpredictable. When absurdity is trivialized and embedded into daily life, it affects social sensitivity and erodes the ability to distinguish fiction from reality, deepening the post-truth era’s disregard for factual accuracy.
More critically, it begins to contaminate the exercise of citizenship, establishing a morally ambiguous zone supported by the normalization of various forms of violence.
Regulation in Brazil: A race against time
In Brazil, Congress is currently debating the regulation of artificial intelligence through Bill 2338/23, approved by the Senate in December 2024 and now under review by the House of Representatives. The proposal brings important advances, including the protection of copyright and a ban on systems generating content involving the abuse or exploitation of children and adolescents—classified as “excessive risk.”
However, the debate faces the powerful lobbying of digital conglomerates, who woo lawmakers to push for weaker regulation aligned with their economic and political interests—proposals that fail to mitigate the societal risks of these tools. Beyond legal and technological questions, the public debate must confront the real challenge of how to halt and hold accountable those responsible for the diffuse harm that AI-generated content inflicts on society—especially videos and audios that defy ethical boundaries and shape individual subjectivity.
*Machine translation proofread by Janaína da Silva.
Autor
Postdoctoral researcher at the Brazilian Institute of Information in Science and Technology (Ibict) in Brazil.
