A student must understand that copying does not develop thinking
Em Guaratinguetá, São Paulo, um estudo com 582 estudantes do ensino médio revelou uma ironia silenciosa da era digital: quanto mais os jovens delegam o pensamento às máquinas, menos conseguem pensar por si mesmos. Quase 70% usam inteligência artificial diária ou semanalmente, mas a grande maioria nunca recebeu orientação pedagógica sobre como fazê-lo com consciência crítica. O que está em jogo não é apenas o desempenho escolar, mas a formação da mente humana numa época em que a conveniência tecnológica pode se tornar uma forma sutil de empobrecimento intelectual.
- Mais de 62% dos estudantes frequentes relatam perda perceptível de memória, e 58% dizem que já não conseguem resolver problemas sozinhos — os músculos cognitivos estão atrofiando em tempo real.
- Quase 68% desses jovens nunca receberam qualquer orientação ética ou pedagógica sobre o uso da IA, navegando às cegas numa tecnologia que remodela silenciosamente sua forma de pensar.
- O padrão dominante de uso — buscar respostas prontas (39%) e gerar textos (25,9%) — foi classificado pelos pesquisadores como 'terceirização cognitiva', a delegação do próprio ato de pensar a uma máquina.
- Educadores como Eucídio Pimenta Arruda alertam para uma segunda crise: a incapacidade crescente de distinguir o que foi produzido por humanos do que foi fabricado por algoritmos, exigindo não apenas habilidades técnicas, mas uma nova forma de ceticismo crítico.
- A resposta proposta vai além de tutoriais de uso: especialistas defendem programas de letramento digital crítico que ensinem os estudantes a questionar os interesses econômicos, os vieses dos dados e as consequências de ceder o pensamento a sistemas automatizados.
Um estudo publicado no Journal Media Critiques acompanhou 582 estudantes do ensino médio em Guaratinguetá, São Paulo, e encontrou uma contradição perturbadora no coração do uso cotidiano da inteligência artificial: as ferramentas funcionam bem demais. Sete em cada dez jovens as utilizam semanal ou diariamente, mas quase 68% nunca receberam qualquer orientação pedagógica sobre como fazê-lo de forma crítica ou ética. Sem essa bússola, os estudantes recorrem à IA como gerações anteriores recorriam à calculadora — mas sem os mesmos limites institucionais.
Os efeitos são mensuráveis. Entre os usuários frequentes, 62,4% relataram perceber enfraquecimento da memória, e 58,1% disseram ter dificuldade para resolver problemas de forma autônoma. Os pesquisadores nomearam esse padrão de 'terceirização cognitiva': a delegação do próprio ato de pensar a uma máquina, especialmente quando o uso se concentra em obter respostas prontas (39%) e gerar textos (25,9%).
Eucídio Pimenta Arruda, professor de educação da Universidade Federal de Minas Gerais, recusa enquadrar o problema como técnico. Para ele, a questão é fundamentalmente humana: copiar não desenvolve o pensamento, não fortalece a inteligência, não constrói o conhecimento necessário para navegar o mundo. Arruda aponta ainda uma crise secundária — a dificuldade crescente de distinguir o que foi criado por humanos do que foi fabricado por algoritmos, impondo um novo fardo cognitivo sobre professores e alunos.
A resposta que ele e outros especialistas defendem é mais ambiciosa do que ensinar a operar ferramentas digitais. Querem programas de letramento digital crítico que levem os estudantes a compreender os interesses econômicos por trás dessas tecnologias, os vieses nos dados que as alimentam e as consequências de usá-las sem julgamento. Arruda vai além: defende a reestruturação da formação docente no Brasil, para que professores de todas as disciplinas possam guiar os alunos a uma relação mais consciente com a tecnologia. A conveniência da IA, sem essa base, transforma-se em armadilha.
A paradox sits at the heart of how teenagers are using artificial intelligence. The tools work beautifully for getting quick answers—too beautifully, perhaps. A study of 582 high school students in Guaratinguetá, São Paulo, published last October in the Journal Media Critiques, found that while these AI systems efficiently solve immediate problems, they appear to be quietly eroding the very skills adolescents need to develop: memory, autonomous reasoning, and intellectual independence.
The numbers tell a stark story. Seven out of ten students in the survey use AI weekly or daily. Yet nearly 68 percent of them have never received any pedagogical guidance on how to use these tools critically or ethically. They reach for AI the way previous generations reached for calculators, but without the same institutional guardrails. When asked why, students cited two dominant reasons: getting direct answers to questions (39 percent) and generating text (25.9 percent). The researchers called this pattern "cognitive outsourcing"—the offloading of thinking itself to a machine.
The consequences are measurable and troubling. Among frequent users, 62.4 percent reported noticing their ability to remember things had weakened. More than half—58.1 percent—said they now struggle to solve problems on their own. These are not marginal effects. They describe a generation watching its own cognitive muscles atrophy in real time, trading the hard work of thinking for the convenience of asking a machine to think instead.
Eucídio Pimenta Arruda, an education professor at the Federal University of Minas Gerais, frames the problem not as a technical one but as fundamentally human. "A student must understand that simply copying something does not develop their thinking, does not strengthen their intelligence, and does not create the knowledge they will need to navigate the world," he said in an interview. The issue is not that AI exists. The issue is that students are using it without understanding what they are doing to themselves in the process.
Arruda identifies a secondary crisis lurking beneath the first: the sheer difficulty of distinguishing truth from fabrication. AI systems now generate documents, images, and videos so convincingly human that teachers and students face a new kind of cognitive burden—not the burden of thinking, but the burden of verification. "The greatest challenge we face today, as educators and learners, is detecting what was made by a machine and what was not," he acknowledged. This is not a problem that can be solved by teaching students to use AI better. It requires something deeper: teaching them to think about the world AI is creating.
The study's authors and education experts like Arruda are calling for something more ambitious than technical training. They want schools to implement programs in critical digital literacy—not to teach students how to operate AI, but to teach them why they should think twice before letting it operate on their behalf. The goal is to transform AI from a tool that replaces human effort into one that amplifies it. This requires students to understand that the companies building these systems have economic interests, that the data these systems learn from can be biased, and that using AI without judgment is not neutral—it is a choice with consequences.
Arruda's vision extends further still. He argues that teacher training in Brazil needs to be restructured entirely. AI is not a problem for computer science classes alone. It touches every subject, every form of knowledge production. Teachers across all disciplines need to understand how these systems work and what they do to learning itself. Only then can they guide students toward a more conscious relationship with the technology. "A student needs to understand that technology can help and strengthen them," Arruda said, "but to benefit from it, they also need to strengthen themselves as thinking, analyzing people—not just analyzing the machine, but analyzing the world." Without that foundation, the convenience of AI becomes a trap.
Citas Notables
A student must understand that simply copying something does not develop their thinking, does not strengthen their intelligence, and does not create the knowledge they will need to navigate the world.— Eucídio Pimenta Arruda, education professor at Federal University of Minas Gerais
The greatest challenge we face today, as educators and learners, is detecting what was made by a machine and what was not.— Eucídio Pimenta Arruda
La Conversación del Hearth Otra perspectiva de la historia
Why does it matter if teenagers outsource their thinking to AI? They're getting answers either way.
Because the brain doesn't work like a filing cabinet. When you struggle to solve a problem, you're not just getting an answer—you're building the neural pathways that let you solve the next problem. Outsource that struggle, and you're not just getting an answer faster. You're atrophying the capacity to think at all.
But couldn't AI be a tool that frees students to think about bigger things?
It could be. But that requires intention. It requires a teacher saying, "Use this to help you, but here's why you need to understand what you're doing." The study found that 68 percent of these kids have never had that conversation. They're just using it.
So the problem is lack of guidance, not the technology itself?
Partly. But there's something deeper. These systems are so good at mimicking human thought that they create an illusion of understanding. A student can generate a perfect essay without learning anything. The technology makes cognitive outsourcing feel like learning.
What does Arruda think teachers should actually do?
He's arguing for something radical: restructure teacher training so that every educator—history, math, literature, everything—understands how AI works and what it does to knowledge production. Not to ban it. To help students use it consciously.
Is that realistic? Teacher training is already overwhelmed.
Probably not realistic in the short term. But Arruda's point is that if we don't do this, we're training a generation that can't think without a machine. That's a bigger problem than any curriculum change.
What would conscious use actually look like?
A student would ask: Why am I using this? What am I learning or not learning by doing so? What biases might be baked into the answer I'm getting? What would I miss if I didn't struggle with this myself? Those questions require a kind of intellectual maturity that doesn't develop by accident.