If AI does that struggle for you, you skip the part that builds your brain.
Em ambos os lados do Atlântico, mais da metade dos adolescentes já recorre à inteligência artificial para realizar tarefas escolares — um salto que, em pouco tempo, dobrou as taxas de adoção anteriores. A maioria ainda usa a tecnologia como apoio, não como substituto completo do próprio esforço, o que oferece algum alento. Mas a velocidade dessa transformação coloca escolas, famílias e educadores diante de uma pergunta que não pode mais ser adiada: o que significa aprender em um mundo onde a máquina pode fazer a lição?
- Em apenas alguns anos, o uso de IA para tarefas escolares dobrou entre adolescentes americanos — de 26% para 54% —, e no Brasil o índice já chega a 59%, sinalizando uma mudança de comportamento que avança mais rápido do que as instituições conseguem acompanhar.
- A maioria dos jovens ainda usa a IA de forma parcial, como suporte ao próprio trabalho, mas um em cada dez estudantes americanos já entrega a tarefa inteira ao chatbot — um padrão que, se se expandir, esvazia o sentido pedagógico das atividades escolares.
- O uso de IA para suporte emocional permanece baixo — 12% nos EUA e 10% no Brasil —, mas a simples disponibilidade dessas ferramentas como substitutos de escuta humana já exige atenção de pais e profissionais de saúde.
- Educadores precisam redesenhar tarefas para exigir julgamento, síntese e pensamento original — habilidades que a IA ainda não replica —, mas essa transformação curricular é complexa e está longe de acontecer na velocidade necessária.
- Pais e estudantes carecem de letramento digital básico sobre como esses sistemas funcionam, onde falham e como usá-los com responsabilidade — uma lacuna que, sem enfrentamento coordenado, transforma a ferramenta em atalho e o aprendizado em simulacro.
Mais da metade dos adolescentes americanos — 54%, segundo pesquisa do Pew Research Center com 1.458 estudantes — já usou inteligência artificial para realizar ao menos parte de suas tarefas escolares. O dado representa uma aceleração notável: em estudo anterior da mesma organização, a taxa era de 26%. No Brasil, a TIC Kids Online Brasil 2025 encontrou número ainda maior: 59% dos jovens entre nove e dezessete anos relataram usar IA para estudar, e 65% a utilizam em alguma atividade cotidiana.
O perfil de uso revela gradações. Um em cada dez estudantes americanos entrega a tarefa inteira ao chatbot; um em cinco pede que a IA faça parte do trabalho; quase um quarto a usa apenas para pequenos trechos. A maioria, portanto, ainda trata a tecnologia como suplemento — não como substituto completo do próprio esforço. Esse padrão é mais saudável do que a terceirização total, mas a tendência aponta em uma única direção.
Um dado oferece algum alívio: o uso de IA para suporte emocional permanece modesto — 12% nos EUA e 10% no Brasil. Ainda assim, a disponibilidade dessas ferramentas como simulacros de escuta terapêutica é motivo de cautela, já que sistemas capazes de imitar empatia sem compreendê-la são substitutos frágeis para a orientação humana real.
O desafio mais urgente recai sobre escolas e famílias. Se a IA resolve em segundos o que antes exigia horas de esforço, a tarefa tradicional perde seu propósito formativo. Educadores precisarão migrar para atividades que demandem julgamento, síntese e pensamento original — uma redesenho curricular que não é simples nem rápido. Ao mesmo tempo, pais e estudantes precisam de letramento sobre como esses sistemas funcionam e onde falham. Sem essa compreensão compartilhada, o risco é que a ferramenta se torne atalho e o aprendizado, apenas aparência.
More than half of American teenagers have turned to artificial intelligence to help with their homework. A survey released today by the Pew Research Center found that 54% of adolescents in the United States have used AI tools to complete at least some portion of their assignments. The researchers interviewed 1,458 students, each accompanied by a parent or guardian, and discovered a spectrum of reliance: one in ten students hand over their entire assignment to ChatGPT or a similar chatbot, one in five ask the AI to handle part of the work, and nearly a quarter report using AI for just a small piece of what they submit.
This marks a sharp acceleration. In an earlier study by the same organization, the adoption rate stood at 26%—less than half of what it is today. The shift happened fast, and it's not confined to the United States. Brazil's TIC Kids Online Brasil 2025 survey, released in October, tells a similar story. There, 59% of young people reported using AI for schoolwork and studying. The Brazilian study reached 2,370 children and adolescents between ages nine and seventeen across the country, surveyed between March and September of 2025. Even broader, 65% of Brazilian youth use AI tools for some daily activity.
One finding offers some reassurance. When it comes to seeking emotional advice or psychological support from these systems, adoption remains modest. Among American teenagers, 12% turn to AI for this purpose. In Brazil, the figure is 10%. This restraint matters because asking a chatbot to serve as a therapist is, by most measures, a poor substitute for actual human guidance. The technology can mimic empathy without understanding it, and teenagers navigating real emotional terrain need more than pattern-matched responses.
But the numbers raise urgent questions about what comes next. Schools and parents are now facing a genuine challenge. If students can offload their assignments to machines with minimal effort, the traditional homework assignment loses its purpose. Teachers will need to rethink what they ask students to do—moving away from tasks that AI can complete in seconds toward work that requires judgment, synthesis, and original thinking. That redesign is not trivial, and it will take time.
Equally pressing is the need for what might be called AI literacy. Teenagers need to understand how these systems work, what they can and cannot do reliably, and where they are likely to fail or mislead. Parents need this knowledge too, because if their children are using these tools regularly and intensely, questions about AI will inevitably surface at home. Right now, many adults lack even basic familiarity with how these systems function, making it hard to guide young people toward responsible use.
The data suggests that most teenagers are not yet treating AI as a complete replacement for their own effort—most are using it as a supplement. That's a healthier pattern than wholesale outsourcing. But the trend is moving in one direction, and the institutions responsible for education have not yet caught up. Schools, teachers, and families are all holding a piece of the puzzle. Without coordinated thinking about what AI means for learning, the risk is that the tool becomes a shortcut rather than a tool, and the work of actual education gets hollowed out.
Citações Notáveis
It is an unwise idea to seek therapy from chatbots— Column author, reflecting on low but concerning emotional support usage
A Conversa do Hearth Outra perspectiva sobre a história
Why does it matter if teenagers are using AI for homework? Isn't that just the natural evolution of how students work?
It matters because homework, at its best, is where learning actually happens—where you struggle with an idea until it clicks. If AI does that struggle for you, you skip the part that builds your brain. The numbers show most kids aren't outsourcing completely yet, but the trend is moving that way.
The emotional support number is surprisingly low—only 10 to 12 percent. Does that suggest teenagers are being wise about the limits of AI?
It could mean that. Or it could mean they haven't thought to try it yet, or they instinctively know that a machine can't actually care about them. Either way, it's the one place where restraint is showing up in the data.
What's the real problem here—is it that students are cheating, or something else?
It's not really about cheating in the traditional sense. It's that the entire purpose of an assignment becomes unclear. If a teacher assigns an essay and the student uses AI to write it, what was the teacher actually trying to teach? That question doesn't have an easy answer, and schools haven't figured it out yet.
So what needs to happen?
Two things at once. Teachers need to redesign what they ask students to do—move toward work that requires real thinking rather than information retrieval. And everyone—students, parents, teachers—needs to understand how these systems actually work, what they're good at, and where they fail. Right now, most people are using them without that knowledge.
Is this a crisis or just a transition?
It's a transition that could become a crisis if nobody acts. The tools are here. The question is whether schools and families can adapt faster than the technology spreads.