The algorithm didn't choose to trap you. Someone did.
Em toda conversa sobre tecnologia, há um momento em que 'o algoritmo' surge como se fosse uma entidade única e malévola — mas um colunista brasileiro decidiu desfazer esse equívoco com precisão filosófica. Algoritmos não são agentes com intenções; são funções de otimização moldadas por decisões humanas e métricas de negócio. A responsabilidade não reside no código, mas nas escolhas de quem o projeta e nos hábitos de quem o alimenta. Compreender essa distinção é o primeiro passo para uma relação mais lúcida com as tecnologias que organizam nossa atenção.
- A expressão 'o algoritmo' virou um atalho perigoso — ela sugere um vilão único quando, na verdade, dezenas de sistemas distintos operam simultaneamente em cada plataforma que usamos.
- A confusão entre ferramenta e agente cria uma armadilha: ao culpar uma entidade imaginária, as pessoas desviam o olhar das decisões humanas reais que priorizam engajamento sobre bem-estar.
- Desenvolvedores e executivos escolheram métricas de retenção como norte — e essas escolhas têm consequências concretas: desinformação amplificada, bolhas de filtro e rolagem compulsiva.
- A saída não é tecnológica, mas educacional: entender como esses sistemas funcionam é o que permite ao usuário agir com consciência em vez de reagir por reflexo.
- O horizonte aponta para feeds personalizados criados pelos próprios usuários, mas a literacia midiática continua sendo a única proteção real contra bolhas — inclusive as que construímos para nós mesmos.
Há um equívoco que se repete em conversas sobre tecnologia: a ideia de que existe 'o algoritmo' — uma força singular e maliciosa escondida dentro dos nossos celulares. Um colunista, ao ouvir alguém falar com absoluta certeza sobre ser controlado por essa entidade monolítica, sentiu a necessidade de desfazer o nó.
A primeira correção é estrutural: não existe um único algoritmo. O que chamamos assim são dezenas de sistemas sobrepostos, cada um tentando prever o que você quer ver a seguir. Instagram, TikTok, YouTube, Spotify, Amazon — todos operam com lógicas semelhantes, mas distintas, ajustando-se em tempo real com base no seu comportamento e nos padrões de pessoas parecidas com você. É aprendizado de máquina em busca de relevância, não curadoria consciente.
Mas o ponto mais difícil vem depois. Algoritmos não têm intenção — porém as pessoas que os constroem, sim. A responsabilidade real não está no código: está nas decisões de negócio que definiram o que o código deveria otimizar. Alguém escolheu que o tempo de permanência na plataforma importa mais do que o bem-estar do usuário. Alguém inseriu comandos manuais para impulsionar certos conteúdos. Essas são escolhas humanas com consequências reais — desinformação reforçada, visões de mundo estreitadas, atenção sequestrada.
Reconhecer que algoritmos são ferramentas, e não agentes, não alivia ninguém da responsabilidade — pelo contrário, a torna mais precisa. A pessoa que se sente controlada pelo algoritmo muitas vezes colaborou ativamente para construir sua própria bolha, fechando-se à contradição e domesticando seus próprios hábitos de consumo.
O futuro pode trazer algum equilíbrio: com a inteligência artificial avançando, usuários poderão criar seus próprios feeds personalizados, escolhendo fontes e formatos de forma deliberada. Mas mesmo nesse cenário, a literacia midiática continua indispensável. Uma bolha construída por você ainda é uma bolha.
There's a moment in every conversation about technology when someone says "the algorithm" as if it's a single, malevolent force—a puppet master pulling strings from inside our phones. A columnist found himself biting his tongue at a bookstore, listening to someone speak with such certainty about being controlled by this monolithic thing, and spent the rest of the afternoon composing arguments he wished he'd made.
The first correction is simple but essential: there is no "the algorithm." The phrase has become shorthand in casual conversation and news coverage alike, a way to describe the personalization systems that feed our social media, music, video, and shopping apps. But the simplification obscures something important. What actually happens is far more intricate—dozens of algorithms and commands layered on top of each other, each one trying to predict what you might want to see next. Instagram, TikTok, YouTube, Facebook, Google, Spotify, Netflix, Amazon, food delivery apps, your bank's mobile application—all of them are running similar but distinct systems, each adjusting in real time based on what you do.
The danger of calling it "the algorithm" is that it suggests a single point of malice, a conscious entity designed to harm you. But algorithms don't want anything. They're optimization functions. They work by observing two things: your own behavior—what you watch, click, like, follow, share—and the patterns of people similar to you. The logic is statistical and expansive: if you engage with knitting content, perhaps you'll also like embroidery. If you share recipes with cilantro, maybe you'll enjoy other cuisines. These aren't individual judgments. They're aggregated patterns that sort people into possible profiles, then adapt based on feedback. It's not curation. It's a machine learning toward relevance.
But here's where the conversation gets harder. Algorithms may lack intention, but the people who build them do not. The real responsibility doesn't live in the code—it lives in the business decisions that shaped the code. Someone decided that engagement metrics matter most. Someone chose to prioritize how long you stay on the platform. Someone built in manual commands to boost certain content. These are human choices, not algorithmic inevitabilities. And they carry consequences: the systems can reinforce misinformation, narrow what you see, keep you scrolling longer than you want to. Understanding this distinction—between the system itself and the goals it was designed to serve—is the only way to think clearly about what's actually happening.
The columnist imagines returning to that bookstore conversation with a second point: knowing that algorithms are tools, not agents, doesn't absolve anyone of responsibility. It actually sharpens it. The person who spoke so confidently about being controlled by algorithms was, in a way, describing her own entrapment—but not the kind she thought. By narrowing what she consumed, by closing herself off from contradiction, she'd built her own filter bubble. She'd domesticated herself. The algorithms had simply made it easier.
What comes next may shift the balance slightly. As artificial intelligence becomes more sophisticated, people may move away from the broad platforms—Instagram, TikTok, the feeds designed by distant companies—toward custom interfaces they build themselves. Imagine creating your own newsletter, your own app, your own feed, pulling from sources you choose and feeding it only what you actually want to read. It would be less like training a pet and more like caring for a tamagotchi: something you nurture deliberately, something you control. But even then, media literacy matters. Even then, the choice to know what you're reading and why remains essential. The alternative is a different kind of bubble, just one you built yourself.
Citas Notables
Algorithms don't want anything. They're optimization functions without agency or intention.— Luiz Henrique Matos, UOL tech columnist
The real responsibility lies not in the code but in the business decisions that shaped it—someone chose to prioritize engagement over accuracy.— Luiz Henrique Matos, UOL tech columnist
La Conversación del Hearth Otra perspectiva de la historia
When you say there's no single algorithm, what do you mean exactly? Isn't there one system running Instagram?
There are many systems running Instagram, layered on top of each other. Each one does something slightly different—one learns from what you like, another from what people like you engage with, another from timing and context. They're all talking to each other, adjusting constantly. Calling it "the algorithm" is like calling a city "the building."
So the real problem isn't the algorithm itself?
The algorithm is a tool. The problem is what it was built to do. Someone decided engagement matters more than accuracy. Someone chose to keep you scrolling. Those are human decisions, not algorithmic ones.
But doesn't that mean we're powerless? If the system is designed to trap us, how do we escape?
You're not powerless, but you have to understand what you're up against. The person in the bookstore thought she was fighting the algorithm by narrowing what she read. She was actually just building her own smaller trap. Real power comes from knowing how these systems work and choosing deliberately what you consume.
Is there a future where this gets better?
Maybe. As AI gets better, people might build their own feeds instead of using platforms. You'd curate your own sources, feed them what you want to see. But that only works if you actually think about what you're doing. Otherwise you just build a prettier cage.
So it comes down to paying attention?
It comes down to understanding that someone made these choices. The algorithm didn't choose to prioritize engagement—a person did. Once you see that, you can start asking better questions about what you're being shown and why.