The search for conversations in AI chatbots

Until recently, to find information online, we would type the subject into the search bar. From the links offered, we would choose those we clicked on. This generated audience traffic and potential monetization for the information-producing media outlets, although platforms like Google didn't always correctly compensate media outlets and other content producers. Now, increasingly, people go directly to artificial intelligence chatbots and formulate their questions. The answer comes organized by AI tools like ChatGPT, DeepSeek, or Gemini. They receive a ready-made text with an overview of the subject, compiled by the chatbot from material available online. Many people are satisfied with this "summary" and stop clicking on the links offered by the other search method. Navigation ceases to be done through clicks and becomes a succession of questions and answers.

The main consequence for the user is that they no longer select the information presented by the links that appear in traditional searches. In this new model of conversational searches in chatbots, the AI ​​tool chooses and prioritizes the information. Most of the time, the user doesn't click on the links to the information sources. This curation ceases to be human and becomes done by the chatbot. The user becomes passive. "It's as if someone were in front of a store telling those outside everything that's inside. The person doesn't need to go in," explains Gustavo Franco, director of the Graphite agency (O Globo, August 24, 2025). Even before, there was already a loss of autonomy, as the prioritization and selection of these links was done by algorithms programmed by the platforms. Autonomy and creativity, which were already limited, become even more reduced and superficial.

Another significant consequence affects content producers. Media outlets, universities, think tanks, and other institutions are having their materials used without the recipient even visiting their websites. Chatbots collect the information and pass it on to the internet user. The intellectual property and audience of these content producers are left unprotected. Without receiving compensation for the use of their materials, how can they continue generating the information that will later be organized and used by AI platforms? In the traditional search model, Google and others connected user searches with the websites that produced the information sought. This allowed for the monetization of clicks, remunerating the websites that had generated the content, even with imperfections. In the new model, websites cease to be clicked and, therefore, have no way to generate revenue from accesses. A new model is created... freeriderThe old and familiar free rider. AI models appropriate the content they find on websites that produce journalistic and academic material and profit from it, but they don't compensate them for it. Furthermore, they prevent the revenue-per-click model from continuing to make newspapers and other producers of quality content viable.

Negotiations and lawsuits are already underway involving news outlets and big tech companies regarding this content remuneration. New York Times and Folha de S. Paul Lawsuits have already been filed against OpenAI, demanding compensation for its content used by ChatGPT. Other proposed solutions include blocking access to protected content. The problem is only beginning. But, as a certain philosopher born in Trier once said, humanity does not set itself problems for which the conditions do not exist.