Meta networks are a breeding ground for online scams, study shows

The Internet and Social Network Studies Laboratory (NetLab), linked to the Federal University of Rio de Janeiro (UFRJ), published Wednesday (Feb. 5) the results of a study about malicious ads on social networks managed by Meta, like Facebook, Instagram, and WhatsApp. The work aimed to increase knowledge about misleading advertising, which bad actors have used to run scams on Brazilian citizens. The results indicate that Meta’s platforms are seen by scammers as fertile ground for the practice.

The study consisted of intensive monitoring from January 10 to 21 of this year. The period coincides with the events following the publication of Normative Instruction 2.219/2024 by Brazil’s Federal Revenue Service. The text made it compulsory for credit card operators and payment institutions to submit certain information on the taxpayers’ transactions every six months. The measure, which came into force on January 1, 2025, triggered a wave of fake news reports claiming that transactions made with instant Pix payments would be taxed. Under pressure from the spread of fake news, the federal government eventually decided to revoke the new rule on January 15.

According to the survey, the outbreak of false stories raised doubts among the population, and fraudsters took advantage of the moment to carry out scams. In all, 151 advertisers were identified who shared 1,770 ads with malicious content. They also mapped 87 fraudulent websites to which users were redirected. In announcing the decision to revoke the ban, the government stated the move was aimed at curbing the circulation of disinformation, among other things. NetLab’s analysis shows that, as far as fraudulent ads were concerned, this objective was not achieved—on the contrary, on Meta’s platforms, this content grew by 35 percent after the pullback.

In numerous cases, ads used simulated pages of public and private institutions. In 40.5 percent of them, the ads were run by advertisers impersonating the federal government. NetLab researchers note that the ads exploit the misinformation surrounding public policies aimed at financial inclusion. It came to their attention that, among the false advertisements, access was offered to both real and made-up government programs.

“The fact that these pages were allowed to run ads on behalf of the government highlights the weaknesses of Meta’s advertiser verification process,” the study notes. Scammers offered services to identify amounts of money users were said to be entitled to receive or advertised the possibility of redeeming money from a given benefit. Users were eventually led to believe they had to pay upfront fees to access these services. Some of these ads also promoted fraudulent guides that instructed consumers to “bypass Pix taxation.”

The researchers believe that the reach of the scams has been maximized by the use of the marketing tools made available by Meta, which allow malicious ads to be boosted and targeted to segmented audiences according to demographic and geographical criteria and user interests. They criticize the lack of transparency in the processing of personal data and see a lack of control and security against misleading advertising, which encourages digital crime on Facebook, Instagram, and WhatsApp.

According to NetLab, by allowing fraudulent ads to be targeted at specific audiences, Meta’s tools give scammers the ability to reach their ideal victims. “There is a vast number of people in the country eager to find an opportunity to move up the social ladder, who need support and public policies from the state to change their lives, which makes people in need a priority target for online scams,” the text reads.

Contacted by Agência Brasil, Meta said that ads that aim to deceive, defraud, or exploit third parties are not allowed on its platforms. “We are always improving our technology to combat suspicious activity. We also recommend people to report any content they believe goes against Facebook’s Community Standards, Instagram’s Community Guidelines, and Meta’s Advertising Standards through the apps themselves,” the note adds.

Artificial intelligence

The study revealed that AI tools were used in 70.3 percent of the 1,244 fraudulent ads. In some videos, deepfakes were found. The tools were widely used to manipulate the image of federal representative Nikolas Ferreira, the main advocate of the campaign to repeal Normative Instruction 2.219/2024. In all, adulterated material involving the congressman appears in 561 ads, 70 percent of which was broadcast after the federal government backed down.

One of them makes use of an original publication shared by Ferreira on his social networks in which he celebrated the repeal of the new rules and claimed that Brazilian workers could be reassured by being able to “use Pix free of the government’s magnifying glass.” The ad, however, manipulates the final part of the video and shows the lawmaker announcing the launch of a measure that would guarantee partial reimbursement of credit card expenses in recent months.

Even though he was the main target of the manipulated footage, Ferreira was not the only public figure to have his image exploited. Finance Minister Fernando Haddad is another figure who appears in different doctored videos, not to mention President Luiz Inácio Lula da Silva, vice-President Geraldo Alckmin, former President Jair Bolsonaro, representatives Eduardo Bolsonaro and Fred Linhares, and journalist William Bonner, among others.

Content moderation

A survey released last year by the Datafolha Institute points out that scams based on Pix payments and bank-issued payment slips are the types of digital crime that generate the most revenue in Brazil, causing BRL 25.5 billion in losses to consumers every year. Also in 2024, identifying the origin of financial scams in Brazil was the focus of a survey carried out by Silverguard, a technology company that offers digital financial protection services.

The study revealed that 79.3 percent of the cases reported by victims through the SOS Golpe platform began on one of Meta’s three platforms, 39 percent on WhatsApp, 22.6 percent on Instagram, and 17.7 percent on Facebook. The results also showed that 7.3 percent of the scams originated on Telegram and 5 percent on one of Google’s two platforms—the search engine or YouTube. The rest were associated with gaming apps, TikTok, email, and others.

A number of studies, NetLab researchers pointed out, have demonstrated the weaknesses in the control of advertising content by Meta platforms. They argue that one of its consequences is people’s lack of credibility in government and public institutions, as the high volume of fraudulent ads reduces users’ ability to identify authentic ads.

Concerns have also been raised about the changes to the rules of the social networks Facebook and Instagram that were recently announced by Meta CEO Mark Zuckerberg. These include the end of the fact-checking program and other changes involving content moderation, such as a reduction in the use of filters that search for posts that violate the terms of use.

NetLab experts point to uncertainties surrounding the scope of these measures. “The lack of specific mention of advertising content moderation by Zuckerberg in his speech makes it unclear whether the changes impact the circulation of fraudulent ads,” the study notes.