Algorithms are increasingly making decisions for people. They are used by companies to select candidates for vacancies, structure content displayed on social networks, support credit analysis, and are even applied in facial recognition systems. At the same time, they promise to make these processes faster...
The way people form their opinions is being profoundly influenced by digital content, a phenomenon that manifests in the daily habit of checking platforms such as Instagram, TikTok, X, YouTube, WhatsApp, Reddit, Snapchat, news portals, search engines, and AIs before even starting the day.
Data provided by DataReportal indicates that by the end of 2025, approximately 185 million individuals were using the internet in Brazil. In October of the same year, the country already registered 150 million social media user accounts, representing over 70% of the total population.
Although people believe they have the autonomy to form their own opinions by consuming diverse sources, they are constantly bombarded by media outlets, media conglomerates, influencers, and content creators with viral topics, sensational calls (clickbaits), repetition of trends, and obvious facts. All of this is mediated by a hidden algorithmic network, increasingly sustained by generative AIs.
The TIC Domiciles 2025 survey revealed that generative AI is already integrated into the routine of 32% of internet users in Brazil, corresponding to about 50 million people aged ten years or older. Unlike journalists or columnists, who have visible subscriptions and trajectories, generative AI appears faceless, without a defined biography, without an explicit position, and without demonstrating emotion, facilitating the production, interpretation, and access to content.
In a scenario where there is a scarcity of time, repertoire, and willingness to verify the truth of what is consumed, generative AI does not only act after the opinion has been established. It intervenes directly in the process of forming that opinion because it explains events, suggests what is considered most relevant, summarizes conflicts, compares different points of view, assesses source credibility, and organizes arguments for discussions.
According to the Reuters Institute, the weekly use of AI chatbots for news increased globally from 7% to 10% in 2026, reaching 13% in Brazil. Researcher Amy Ross Arguedas, in her report, points out that these systems are used not only to get news but also to question, summarize, simplify, and judge information.
Reuters Institute data shows that among those who use chatbots for news, 42% use them to ask follow-up questions, 35% to find current news, 34% to summarize articles, 30% to facilitate understanding, and 33% to assess the reliability of a source. The primary reasons cited by users are the search for greater depth or clarification (mentioned by 42%) and the need to obtain information more quickly (cited by 39%).
It is undeniable that technological advancement cannot be rejected, nor can the positive potential of generative AIs to boost science, research, productive processes, development, and innovation. However, the central debate lies in how capable we are of challenging answers before accepting them as true, and how prepared we are to accept that generative AI can convert information into an invisible input of answers that the public consumes without recognizing its origin.
The real danger, therefore, does not lie solely in machine failures, but rather in the normalization of a society that is beginning to delegate the function of doubting, interpreting, and gradually, even the intrinsic capacity to form opinions.
