OpenAI has initiated a pilot program to introduce advertising within ChatGPT to users located in Brazil, specifically focusing on free and Go accounts. These ads will be inserted during interactions, always marked as sponsored, and will not impact the responses provided by the service.
Brazil joins other countries in this effort, assisting OpenAI in developing the advertising experience on ChatGPT. It is important to note that paid accounts, such as Plus, Pro, Business, Enterprise, and Education, will not be affected by this pilot implementation.
This move was expected, given that OpenAI had already begun testing ads on ChatGPT in the United States in February of this year, with presumed plans to expand to other nations over time.
In the Brazilian context, advertising tests will occur on the Free and Go versions of ChatGPT, following the model adopted in the United States. Additionally, the display of these ads will be restricted to users over 18 years of age.
OpenAI emphasizes the importance of Brazil in this process, seeking to learn from a market that not only follows changes but also actively contributes to driving them, using this local dialogue to create a relevant experience for both individuals and businesses of various sizes. Dave Dugan, Head of Global Ads Solutions at OpenAI, reinforced this objective.
The ads are clearly identified as sponsored. Although they appear in the middle of the ChatGPT response flow, they are highlighted to avoid any confusion with the service's original content. OpenAI clarifies that user conversations, personal data, and memories will not be shared with advertisers; they will only have access to general campaign performance metrics, such as clicks and views.
The company guarantees that the advertising materials will not interfere with the responses generated by the system. In addition to Brazil and the United States, ad tests are already taking place in markets such as Canada, Australia, New Zealand, the United Kingdom, Japan, and South Korea. Mexico is also scheduled to participate in an advertising pilot program soon.
Social media users have been targeted by scams orchestrated through fake advertisements posing as the Desenrola Brasil program. These fraudulent pages promise to simplify debt negotiation, but their real goal is to collect victims' personal data or redirect them to contacts controlled by criminals.
According to a survey conducted by Agência Brasil, based on data from UFRJ's Netlab, 544 fraudulent ads were identified across Meta platforms between April and June, including Facebook, Instagram, WhatsApp, Threads, and Messenger.
The study conducted by the Laboratory of Internet and Social Network Studies (Netlab), affiliated with the Federal University of Rio de Janeiro (UFRJ), found that 48 pages were responsible for disseminating these deceptive campaigns. Of this total, 278 ads mentioned the official name of Desenrola Brasil, while another 264 promoted a fictitious program called Libera Brasil. In some cases, a page combined both names to simulate authentic communication.
Desenrola Brasil, established in 2023, was created to help consumers renegotiate their debts. Netlab observed that the circulation of these fake ads increased after the launch of the program's new version in May.
The methods employed by scammers include the improper use of government or recognized brand visual identity and names, the use of links mimicking official websites for information theft, the promise of immediate financial resolutions, sending messages that induce urgency, and creating content using artificial intelligence.
Marie Santini, director of Netlab, told Agência Brasil that the difficulty lies in the fact that genuine and fake ads are displayed side-by-side on the same social media platforms, without adequate filters or security systems.
When users interact with these ads, they may be redirected to pages set up by the scammers. In these environments, the intention is to obtain confidential and personal data or direct the victim to conversations via WhatsApp.
The survey revealed that 72% of the analyzed ads improperly used names or images belonging to the government and private brands. Additionally, 19.1% led to URLs imitating official channels, and 38.1% contained AI-generated material. Researchers indicate that these tactics make the scams more believable and complicate fraud detection.
In response, Meta communicated that it implements actions to combat digital fraud, employing artificial intelligence, large language models (LLMs), facial recognition, and human analysis. The company also assured the removal of any content that violates its guidelines, stating: 'We use a combination of technology and human review to identify and remove violating content.'
The study demonstrates that criminals persist in exploiting established programs to mislead users. Therefore, before sharing data or accessing any link received through ads, it is essential to verify whether the communication comes from an official channel.
