Marcelo Alves
How Brazilian disinformation industries and their practices of hate speech, and antidemocratic movements are monetized? This broad question guides my new research agenda dedicated to understanding the economic sustainability of false information and content junk on digital platforms. Particularly, I seek to analyze how big tech, ad tech networks and its programmatic advertising infrastructures directly incentivize and reward the production of disinformation at large and provide means of organization of far-right and extremist movements that challenge the democratic order. More dramatically, the economic and data colonialism from big tech provide different challenges for understanding and ascribing potential policies to regulate the problem.
History
This relatively new research topic has been in the spotlight ever since the widely known case of Macedonian teenagers that leveraged political websites and Google’s AdSense technologies to profit over 2016 Election fake news. In Brazil, there is a similar example which directly implicates Google Latin America division in providing training and technical guidelines to far-right bloggers. At the meetings, held at Google headquarters, anti-leftists bloggers were taught how to “optimize” political content for profit on Google platforms, in some cases earning up to 25.000 BRL monthly (roughly 5.000 USD). The false headlines were tailored to outrage and provoke the sense of an urgent revelation, such as: “Dilma Rousseff [former president] caught personally commanding R$48 million kickback scheme”.
The year of 2016 was a major laboratory for far-right digital tactics. At that moment, the country was appalled by the Car-Wash Operation – a lawfare procedure largely designed to criminalize politics and imprison leftist leaders; and the Parliament voted to impeach President Dilma Rousseff from office for technical charges of manipulating the federal budget. Research demonstrated that the criminalization of politics was crucial to the mobilization and organization of a large, multiplatform and coordinated ecosystem of far-right digital media. Spearheaded by Jair Messias Bolsonaro, this network of activists and rising political influencers successfully and systematically capture the public attention, earning stellar digital metrics compared only to the largest entertainment and soccer celebrities.
Context
Since then, the problem grew larger and much more complicated. The Global Disinformation Index (GDI) found that 20.000 websites flagged for disinformation profited roughly a quarter billion dollars. In Brazil, an investigation led by the Supreme Court revealed a disinformation industry operating on Youtube that amassed more than 6.8 million reais, between 2018 and 2022. The largest channel of the list is Folha Politica, a far-right fake news operation that tallied more than 1 billion views anda profited approximately BRL 2.59 million reais only on Youtube.
The Sleeping Giants movement has been crucial to mobilize the public attention to the brands and advertisers that help to monetize junk news and hate speech. The Brazilian branch of the group has conducted 47 demonetization campaigns, resulting in roughly BRL 76 million interrupted advertising reais from 1054 companies.
A part of this industry is also funded indirectly by the state. The federal government Secretary of Social Communication placed more than 2.65 million ads on junk news and problematic websites, such as gambling, pornography and fake news. Among them, the congressional investigation demonstrated that Allan dos Santos, convicted and fugitive from Justice on charges of apology for antidemocratic movements, has profited from federal resources. Although the publicly known data is scarce and incomplete, I would argue that funding from federal government and state companies is an important but secondary source of revenues for the Brazilian disinformation industry.
Early results of ongoing research
There is a huge gap in the academic literature in the topics of disinformation studies or platforms and society regarding the role of ad techs and the digital infrastructure in providing economic incentives for disinformation industries or campaigns. Most of the work is very recent and led by computer scientists, focused mostly on accountability purposes. Braun and Eklund (2019) demonstrated how the programmatic advertising exchange and digital media infrastructures provide lucrative businesses to disinformation industries. Their research shows how ad tech stakeholders face the problem largely from a technological or marketing perspective, i.e. as a “brand safety issue”, disregarding the democratic implications.
My new research aims to understand the financial subsidies to the Brazilian far-right disinformation industry. This research is still in its very initial phases of literature review and construction of extraction instruments. I conducted an early study to detect ads on the TerraBrasilNoticias. The website is a major hub of far-right disinformation pro-Bolsonaro, even using a paraphrase of Bolsonaro’s catchphrase “God above all and everyone”. Their administrators reportedly profit between BRL 12 and 6 thousand reais monthly from programmatic advertising, mostly copying pieces and altering headlines from journalistic outlets.
Although Google’s DoubleClick is the single major ad exchange (Papadogiannakis et al., 2022) both in professional journalism (96%) and junk news websites (80,8%), there is a prevalence of poor quality ads in lesser known ad techs. For example, study shows Taboola owns the biggest share of problematic ads, a content category encompassing financial frauds, gambling, malwares, content farms and unapproved miracle tonics (Zeng et al, 2020). Frequently, smaller ad techs have looser governance policies and smaller incentives to enforce content moderation on websites they serve advertising.
I collected approximately 27 thousand ads provided by MGID on TerraBrasilNoticias story pages, georeferencing the browser in Brazil. MGID is a growing ad tech company that recently opened headquarters in Latin America. According to their website, the service is tailored to insert native advertising using artificial intelligence to closely resemble editorial content. More importantly, they have even less strict enforcement of content moderation than Google AdSense, which impacts both ends of the programmatic ad exchange. At the supply side, MGID has poor quality control on the types of ads served by its systems, allowing several types of frauds; and at the demand side, the company is negligent in filtering hate speech or false information.
Early findings demonstrated that none of the ads collected from MGID’s at TerraBrasilNoticas represents major national or multinational enterprises. Actually, content analysis indicates that the largest percentage of them are content farms, i.e. false websites or blogs that create sensational or clickbait headlines to generate traffic and revenue. The second most important category is financial fraud and gambling, most of them adopting storytelling style of individual cases of people who became rich following an investment strategy or buying cryptocurrencies or NFTs, such as: “Woman earns over $480,000 working from home and keeps secret from her husband for two years”, “Blockchain Cuties – Play2Earn NFT Game! Fights Breeding Trade!”, and “Take that betting tactic. My secret weapon for victories”.
The results show how platform scams (Grohmann et al., 2022) are pervasive in the digital attention economy. Frauds and deception are largely manipulated algorithmically to blur frontiers between journalism and advertising in artificial intelligence generated native ads that not only fund disinformation content, but also are fake, misleading and abusive on their own. A minor but worrisome percentage of the ads are weight loss tonics unapproved by the national health agency. We found that these native ads point to hot site urls that copy the layout of professional journalism vehicles, telling fake stories of miraculous slimming. The Brazilian fact-checking agency Aos Fatos reported that some of these products were prohibited by the regulatory agency. A few months after the data collection, our urls to these ads became inactive and the whole websites were taken down.
Scams, fakes and frauds are pervasive in the digital economy (Poster, 2022). In Brazil, they are a part of a digital infrastructure of deception that perform three functions: a) funding disinformation websites; b) creating false stories, reviews and testimonies to sell financial pyramids, illegal health products, or c) arbitraging digital traffic though clickbaits and fake news. Further research is necessary to study interconnections between platform power, data colonialism and disinformation studies in the Global South, especially in face of the lenience and complicity of the profit-making big tech companies.
The Brazilian case challenges the debate on policy and potential regulations since there is an enforcement lenience and cultural miscomprehension of local political culture, institutions and regulations. Recent ProPublica investigation of global funding of Covid disinformation found that Eastern Europe, Africa and Latin America still receive ads from major brands on unreliable websites that violate Google policy: “ProPublica’s examination showed that ads from Google are more likely to appear on misleading articles and websites that are in languages other than English”.
From the outset of this research, I would argue that the infrastructure of digital platforms and the monopolization of the attention economy facilitate, fund and empower far-right media ecosystems, at the same time that create precarious conditions of click-working, play-to-earn and pay-per-watch schemes. Advancing the fight against disinformation industries in Brazil is not only a matter of building more precise artificial intelligence tailored to portuguese language and national cultural specificities, but actually would demand facing the very monopolization of digital infrastructures that fragment the public sphere and monetize increasingly hateful and antidemocratic speech.
Marcelo Alves is Professor of Media Studies in Pontifical Catholic University of Rio de Janeiro (PUC-Rio)
This op-ed is part of the project “Global Democracy Frontliners: Transnational Research Coalition for Tech Accountability and Democratic Innovations Centering Communities in the Margins”, funded by Luminate.