Her head was underwater, submerged in content designed to drown her.
Uma geração de jovens europeus cresce imersa num ambiente digital que, segundo a Amnistia Internacional, os empurra sistematicamente para conteúdos de automutilação e suicídio — não por acaso, mas por design algorítmico. Um estudo com mil jovens franceses revelou que mais de metade associa o uso do TikTok à depressão e à baixa autoestima, enquanto testes controlados mostraram que a plataforma inunda perfis fictícios de adolescentes com conteúdos nocivos sem qualquer estímulo prévio. É uma questão antiga com uma forma nova: até onde vai a responsabilidade coletiva de proteger os mais vulneráveis dos ambientes que construímos?
- O algoritmo do TikTok promoveu ativamente conteúdos de automutilação e suicídio a perfis simulados de jovens de 13 a 15 anos, sem qualquer interação prévia que o justificasse.
- 58% dos jovens inquiridos relatam sentir-se deprimidos e com a autoestima destruída como consequência direta do consumo da plataforma.
- Testemunhos como o de Maëlle, 18 anos, revelam jovens que deixaram de conseguir ir à escola ou funcionar no dia a dia, submersos numa espiral de conteúdos nocivos.
- Os 25 estados-membros da União Europeia emitiram uma declaração conjunta sobre proteção de menores online, mas permanecem divididos quanto a medidas concretas, incluindo eventuais restrições de acesso.
- A comparação com espaços físicos torna-se inevitável: plataformas digitais operam com uma liberdade e impunidade que seriam inaceitáveis em qualquer outro contexto regulado.
A Amnistia Internacional em França passou meses a ouvir mil jovens entre os treze e os vinte e cinco anos sobre a sua relação com o TikTok. Os resultados, divulgados esta semana, combinam dados quantitativos com testemunhos diretos e testes de navegação controlados. O número mais perturbador: 58% dos jovens inquiridos afirmam que o conteúdo da plataforma os deixou deprimidos e com a autoestima corroída.
Mas o comportamento do algoritmo em laboratório revelou-se ainda mais inquietante. Quando investigadores criaram perfis fictícios de jovens entre os 13 e os 15 anos e simplesmente navegaram na plataforma, foram inundados com vídeos que promoviam a automutilação e o suicídio — sem qualquer histórico de pesquisa ou interação prévia que o justificasse. O sistema decidiu, por conta própria, que eram esses os conteúdos adequados para um adolescente.
Maëlle, 18 anos, descreveu ao pormenor o que isto significa vivido por dentro: houve um período em que mal conseguia ir à escola, em que os dias passavam sem energia nem propósito, horas a fio no TikTok. Para descrever a sensação, usou uma imagem simples — tinha a cabeça debaixo de água.
A União Europeia emitiu uma declaração conjunta sobre proteção de menores online, mas os 25 estados-membros permanecem divididos quanto às medidas a adotar, nomeadamente sobre eventuais restrições de acesso às plataformas. O que parece consensual é que a literacia digital e a moderação de conteúdos precisam de ganhar substância real. Regulamos os espaços físicos com rigor e consequências. As plataformas digitais continuam a operar numa liberdade que, à luz destes dados, se torna cada vez mais difícil de justificar.
Amnesty International in France spent months listening to a thousand young people between thirteen and twenty-five years old talk about their relationship with TikTok. The researchers released their findings this week, and what emerged was both quantitative and deeply personal—surveys paired with direct testimony, supplemented by the organization's own navigation tests conducted under controlled conditions. The numbers were stark: fifty-eight percent of the young people surveyed reported that TikTok content had left them feeling depressed, their self-esteem eroded. But the algorithm's behavior in the lab was perhaps more disturbing still.
When Amnesty researchers created fake profiles posing as thirteen- to fifteen-year-olds and simply scrolled, the platform inundated them with videos promoting self-harm and suicide. No search history prompted this. No engagement with similar content had trained the algorithm in that direction. The feed simply arrived, unbidden and relentless, as if the system had decided these were the videos a young teenager should see.
Maëlle, eighteen years old, gave the researchers a window into what this looks like from the inside. She described a period of her life when she barely attended school, when she moved through her days without energy or purpose, spending hours on TikTok. To capture the feeling, she used a single phrase: her head was underwater. She was submerged. The mental health impact on young people is the report's central concern, and it is why the researchers are calling on authorities to act.
The European Union's member states recently issued a joint statement affirming their commitment to protecting minors online. Yet the twenty-five nations remain deeply fractured about what that protection should look like—particularly whether it should include restricting access to these platforms altogether. Some want bans. Others resist them. What seems clear is that serious digital literacy education is overdue, and that content moderation standards need to be defined with real teeth. The comparison is unavoidable: we regulate physical spaces with care and consequence. Digital platforms, by contrast, operate with a freedom that seems incomprehensible. We have constructed a system where it is easier to harm young people online than to face accountability for doing so. The platforms continue their work while thousands of young people, like Maëlle, struggle to keep their heads above water.
Citas Notables
I felt like my head was underwater—without strength to do anything.— Maëlle, 18, describing her experience during heavy TikTok use
La Conversación del Hearth Otra perspectiva de la historia
Why does Amnesty's testing matter more than just asking young people how they feel?
Because the algorithm doesn't care what you ask it to do. When researchers created fake teenage profiles and did nothing—no searches, no clicks, no engagement—the system still served them content about self-harm. That's not a bug. That's what the system is designed to do.
But surely TikTok would say they have safeguards, that they're trying to prevent this?
They say that. And maybe they try. But the evidence here is that their safeguards failed completely when tested. A brand-new account, no history, no signal—and within hours, the algorithm is recommending ways to hurt yourself.
Maëlle's phrase about her head being underwater—that's not just sadness, is it?
No. It's describing a specific kind of paralysis. She couldn't go to school. She couldn't move. She was drowning in content designed to keep her scrolling, and the content itself was designed to make her feel worse. The two things fed each other.
The EU member states can't agree on what to do. What would actually work?
You need both things: real education so young people understand what's happening to them, and real consequences for platforms that don't moderate. Right now, we treat a physical store that sells harmful products to children as a crime. We treat a digital platform that does the same thing as innovation.
Is the answer to ban TikTok?
That's the question dividing Europe. But even if you ban it, the next platform will use the same algorithm. The real answer is making the platforms answer for what they do.