In Brazil, Meta's platforms host job ads linked to human trafficking recruitment – Núcleo Jornalismo

In January alone, Núcleo identified 28 ads matching trafficking-recruitment patterns Brazil's Foreign Ministry has flagged.
A cleaning job in the United States paying up to $75 an hour, with no English required, no experience needed, flexible hours, plus housing and cash payments. The offer is aimed squarely at immigrants and newcomers.
Ads like these circulate on Facebook, Instagram, and Threads, targeting Brazilian audiences. While Meta enforces strict transparency rules for job ads in places like the United Kingdom, United States and Canada, in Brazil these offers exist in a regulatory blind spot. Dressed up as the opportunity of a lifetime, the pitch carries clear signs of fraud that Brazil’s Ministry of Foreign Affairs has already flagged as the opening move in human trafficking schemes.
Between January 1-31, 2026, Nucleo’s investigation identified at least 28 job ads in Meta’s Ad Library showing signs of recruitment for exploitation — nearly one a day. The same cleaning job ad was replicated 16 times on the platform. Another 12 ads followed the same script, this time targeting healthcare workers.
On February 24, the Ministry issued an alert about signs of recruitment for human trafficking in job ads in Southeast Asia. It found that “offers, specifically targeting Brazilians, include seemingly attractive salaries, commissions for ‘assets’ sold, and coverage of airfare.”
Data obtained through Brazil’s FOIA requests show that, among 63 potential victims of human trafficking recorded by consular posts in 2024, 40 were classified as workers in conditions analogous to slavery that were working on digital betting platforms in some capacity.
Geographically, the problem is more concentrated in Southeast Asia, with the Philippines (20 cases), Laos (11), and Cambodia (7) leading the rankings. Although the region is emerging as a hotspot for human trafficking—particularly targeting young people with IT skills— this is neither a new problem nor one exclusive to Asia.
On March 4, the Embassy of Brazil in Croatia also issued a consular alert reporting an increase in the number of Brazilians lured by job offers who end up as victims of human trafficking networks for sexual exploitation in the European country.
Finding these ads in Meta’s Ad Library, however, is no easy task, as the platform does not provide a job listings category in Brazil.
In practice, this means job ads are mixed in with listings for products, services and all kinds of other advertising, and they become inactive once their timeline to run ends—unlike political ads, which have their own category and must be stored for up to seven years for public access. The absence of a dedicated job ads category in Brazil makes auditing more difficult for researchers and journalists.
While finding these ads for auditing purposes is difficult, being found by an ad is precisely the service Meta offers to advertisers, who can target campaigns based on age, gender and location. Other factors are based on user behavior: what they search for, when and how. All of this, of course, takes place under trade secrecy, as algorithms are considered strategic assets by those platforms .
“Regarding the algorithm, we can already say that there is a form of gender-based violence,” says Julice Salvagni, a labor sociologist and coordinator of Fairwork Brasil, a project based at the Oxford Internet Institute that studies work on digital platforms. “For women, lower-paying jobs are offered, which shows that gender inequality is also being perpetuated in this sphere,” she adds.
The company does not disclose the demographic distribution of ads, which makes it hard to verify what audiences the content is being targeted at. Targeted job ads in Brazil are illegal, but because the platform does not distinguish between product ads and job listings, it is able to circumvent the law that bans such targeting.
Rodrigo Castilho, a prosecutor at Brazil’s Public Ministry of Labor (MPT), notes that Brazilian labor laws prohibits the publication of job offers targeted by sex, age, or race. “The main problem is that these laws do not apply to a company headquartered outside Brazil,” he explains.
Among those ads, fraudulent job offers can lead to the recruitment of Brazilians by human trafficking networks. “It is not difficult to place this kind of ad on some of the world’s most widely used platforms, such as those operated by Meta, which does not exercise proper oversight. These ads face no barriers in reaching users,” says David Nemer, a technology anthropologist and professor at the University of Virginia.
For David Nemer, digital platforms are more compliant in countries in North America or Europe than in Global South nations like Brazil. In fact, ad transparency policies are not the same worldwide. While Brazil debates the regulation of social media, Meta faces lawsuits around the world that end up forcing the company to implement transparency policies.
Although this dynamic of scams and fraud is not entirely new, social media platforms intensify and accelerate the problem: 86% of the Brazilian population has internet access, while 81% use social media. Overall, 19% said they use the internet to search for jobs.
Social media platforms create a conducive environment for advertising job openings, but their architecture also provides a layer of protection for recruiters that the physical world did not allow, says Natália Suzuki, Education and Public Policy Manager at Repórter Brasil. According to her, the digital environment makes the process “more fluid,” as those responsible for exploitation can operate from other regions.
“The person may never show up, may never reveal their identity. Catching them is much harder than identifying a ‘gato’ [exploitative labor recruiter] or a concrete illegal labor intermediary,” Suzuki said.
Since it is not possible to search by a job ads category in Brazil, the investigation had to search across the entire Meta Ad Library. Throughout January 2026, the keyword “call center” returned 77,881 ads. The keyword “model” generated another 61,180 results, while the similar term “model contract” yielded 29,525 ads. Combined, these terms total 168,586 ads (all these terms were searched originally in Portuguese).
Running the same search via the API in countries where Meta complies with stricter transparency rules returns a substantially lower volume of results. For the United Kingdom, a similar search yielded 657 ads. For Canada, 237.
Even if a “job” category comes along, this may not be enough to ensure transparency, as Meta avoids responsibility, saying that properly categorizing ads is the advertiser’s duty, not the company’s, according to Bruno Mattos, project coordinator at NetLab, a research lab at the Federal University of Rio de Janeiro (UFRJ).
“Platforms place a great deal of responsibility on advertisers, but we cannot lose sight of the fact that they are profiting from these ads and must be accountable for the various stages of verification, advertiser approval, and the entire advertising lifecycle,” he said.
Brazil’s Ministry of Labor and Employment has not responded to our request for comment.
Meta said that “activities intended to deceive, defraud, or exploit others are not allowed on our platforms, and we are continually improving our technology to combat suspicious activity. We also encourage people to report any content they believe violates the Facebook Community Standards, the Instagram Community Guidelines, or Meta Platforms’s Advertising Standards through the apps themselves.”
To explore the job ads ecosystem, we accessed Meta’s Ad Library through the Graph API using a Python script. We began by searching for broad terms like “call center,” “customer service,” “cleaner,” and “model,” which returned a huge number of results. Most of them, however, were AI-generated content that cluttered the dataset and made it harder to analyze the real ads.
The report refined the search using more specific terms, focusing on “Asia,” “Cambodia,” and “Thailand,” in line with consular warnings and recent news reports that highlight this region as a hotspot for human trafficking.
By narrowing the search, we identified 1,967 ads. Through textual analysis, we flagged 944 as fraudulent, though not all were related to overseas job offers or showed signs of recruitment schemes.
The consular warning was used as a guide to categorize ads showing signs of recruitment. The result was 28 ads promoting overseas jobs for low-skilled workers, offering accommodations, requiring little to no foreign language proficiency, and showing a preference for immigrants.
Judge's decision contained a command to improve a paragraph and the AI's rewritten text
Of 4,000 vehicles analyzed, only 7% block AI company agents, according to a survey by The Protocol Gap project.
Tech giants operating in the country assembled a robust lobbying apparatus that successfully fended off civil society groups and the federal government to protect their interests in Congress.
Unreliable AI detection tools are used as a form of worker surveillance amid a vacuum of regulation and ethical discussion
Sem regulação, ofertas de emprego no Facebook, no Instagram e no Threads violam leis trabalhistas e prometem condições semelhantes às que foram usadas para aliciar brasileiros que se tornaram vítimas de tráfico de pessoas
Investigação do Núcleo mapeou mais de 20 canais que utilizam inteligência artificial e o conceito de estoicismo para criar uma fábrica de conteúdos misóginos, sintéticos e monetizados pela plataforma
Queremos saber sua opinião e o que você tem a dizer. Você também pode nos encontrar em nucleo.jor.br/comunidade.
Comentários estão sujeitos à moderação de nossa equipe. Críticas são bem-vindas, mas xingamentos, ameaças e discurso de ódio serão removidos, assim como piadas racistas, homofóbicas ou transfóbicas.

source