Presented initially at II COCA, Universitat de Girona, January 2022 as part of the panel on Promise and Failure of Infrastructures
[*] Raquel Rennó, PhD. Article 19 and Juliana Novaes, MA candidate at Maastrich University and Article 19 Internet of Rights Fellow.
Since the Cambridge Analytica scandal, a growing body of research shows how digital platforms are exploited by malicious actors causing interference in democratic processes worldwide. Special emphasis has been given to the content of disinformation and misinformation, but there is not much discussion about the role of Internet infrastructure in this context. Combating misinformation is not only a matter of what information is being spread but how it is spread, since infrastructures are also part of the discursive constructions. This study focuses on the Brazilian context to illustrate the case. Mobile Internet, the main way of accessing the Internet in Brazil, in its prepaid version, is offered frequently with “unlimited access” to the most popular applications, a practice called zero rating where specific content is tagged and not counted in the data cap. This contributed significantly to the extreme concentration in the use of a few applications, like Whatsapp, the most used mobile messenger in Brazil, especially among the lower socioeconomic segments of the population. By limiting the data use through the imposition of data caps and concentrating access to services that do not have to comply with any media regulation, this configuration has been easily exploited by campaigners and disinformation perpetrators. The messaging service, especially Whatsapp, has been weaponized and used as a media source for political information that is not subject to public transparency or scrutiny, since it is not an official media channel. This study aims to analyze and make visible the communication architecture from the perspective of Star´s critical and relational infrastructure, detailing how this specific communication network was developed from the data traffic layer to the social tissue and how different processes of local infrastructure public regulation and centralized private sector strategies serves as basis for the political misinformation campaigning in the 2018´s Brazilian presidential elections.
The politics of Internet infrastructure in Brazil
The Internet is part of telecommunication infrastructure, and as with other telecommunications services like broadcasting and satellite, it requires the use of radio spectrum. This is a natural resource that facilitates the wireless transmission of information over long distances, and its frequencies are studied, managed and defined globally by the International Telecommunications Union (ITU). Similar to other infrastructure systems such as transportation systems, water supplies, and electricity grids, the Internet infrastructure planning and expansion is mostly seen through an socioeconomic development lens; its deployment is also mediated through public-private partnerships and the frequent mismatch between large-scale planning and the local needs of communities, a mismatch that often creates one-size-fits-all projects that, in most cases, create more problems than solutions for people and communities.
Internet access, in general, depends on the interconnection of a series of networks that starts from transnational and undersea infrastructure, passes through national backbones and backhaul, and extends to what is called “the last mile”, the local networks that finally connect the Internet user. Just to be deployed, a broadband connection requires overcoming topographic barriers, coverage over large extents of land, structural support such as cellular towers and roads, and investments in technology that allow for fast transmission of data. When taking into account all of the technological and infrastructural requirements associated with broadband expansion, these services involve significant financial investments (OECD, 2018). In an industry-driven market, where the private sector is responsible for developing and deploying infrastructure, these activities become a matter of business investment and decisions of where and how this investment will occur become a matter of maximal profit. The Internet infrastructure network in Brazil, much like in most parts of the world and especially in the Global South (or Global Peripheries), is the result of the lack of interest from commercial Internet service providers to offer service, whether affordable or at all, to rural, remote and low-income communities (Knight et al., 2016). This is clearly visible today, where the lack of connectivity is a regional and rural issue:
Since access to fixed broadband, the kind of service that is “always on a fixed location” and is provided by DSL (Internet transmitted via phone lines), fiber optics cable and other technologies, is largely unavailable or unaffordable, a significant part of the population in these unserved territories rely on mobile connectivity for their daily activities. A mobile phone is the main point of access to the Internet in the Global South, specifically remote and rural areas. This is due to issues of availability, as many areas in the Global South are not covered by fixed broadband service, and affordability, since mobile networks offer the option of pay-per-use plans for smaller data packages.
According to data from the Regional Center for Studies on the Development of the Information Society (2021), 58% of Brazilians access the Internet exclusively through a mobile connection, reaching 91% in the lower economic segments.
Given these conditions, in which fixed broadband Internet connectivity, where there is generally no data cap, is unavailable, insufficient, or unaffordable, users from lower socioeconomic backgrounds develop a dependency on zero-rated services and applications included in mobile Internet plans. Zero-rating is a term that refers to a specific practice in telecommunications, whereby consumers are offered plans that enable the use of certain services and applications without being charged as part of the data cap. Normally, this offer is provided as part of prepaid plans, where the user gets a limited amount of data to use in a certain period of time. Zero-rating plans are usually the outcome of negotiations between technology companies that develop applications and services and network operators that offer mobile Internet plans to users. Given the scale of this negotiation, these zero-rating agreements are generally made between multinational technology companies and network operators that have already achieved market dominance. In a recent study on Internet users in Brazil, 40% of low-income individuals reported having their connection recently restricted exclusively to zero-rated applications due to the impossibility of buying more mobile data. Among them, 80% mentioned they would prefer to have the possibility to have access to other platforms and websites on the Internet, rather than being limited to one zero-rated application (Instituto Brasileiro de Defesa do Consumidor & Instituto Locomotiva, 2021).
From an economic point of view, some may see this practice as an opportunity for lower income segments of the population to start using the Internet, increasing the connectivity for all (Basri, 2019). However, local human rights experts say that it has become a way for large mobile network operators to control infrastructure demand and compensate for poor private sector infrastructure investment in rural and low-income neighborhoods. According to Canabarro et al. (2016), “the strategy gains attractiveness with the limitation of data franchises on the mobile Internet, created under the argument of the scarcity of sufficient physical infrastructure to meet the demands of the evolution of the network and the growing volume of users and traffic data” (Canabarro et al. 2016). Lefebre (2017) adds to this idea, saying that what was originally supposed to be a temporary relief has become a permanent fix for the lack of infrastructure. She states that it “is not possible to address the effects of the commercial practice of zero-rating and sponsored access, especially in developing countries, without taking into consideration the insufficient telecommunications infrastructure that supports the Internet connection service”.
As mentioned before, the use and management of spectrum are under the authority of states, who can then license the use of certain frequencies to telecommunication companies via auctions and other licensing processes. This licensing agreement normally includes conditions for infrastructure expansion, meaning that the company that gets the license in an auction also has to deliver infrastructure to areas where it is needed. By using prepaid plans combined with zero-rated applications, the private sector can avoid the costs of infrastructure deployment in areas where there is not sufficient demand to justify commercial investment. For general development indexes and general government concerns, these users are counted as being “connected” despite having very limited access to information. WhatsApp and Facebook Messenger (both belonging to the same company, Meta) are the most commonly zero-rated applications and are often implemented so that the zero-rated versions are text-only or contain very limited imagery. As such, they take up relatively little bandwidth and do not place much pressure on network infrastructure. There are different types of prepaid plans, including the ones that include a relatively high data cap, up to 95GB. The charging model in itself is not a problem, but it becomes a problem when insufficient data caps, combined with zero-rated applications, is the primary or only way that people and communities from low socioeconomic backgrounds can access the Internet. Official information about the amount of data used in rural and suburban areas and specific income segments should be made public so it can be properly tracked and studied.
Over time, the increase in the use of smartphones as the main or sole point of Internet access gives disproportionate power to one type of service provider, in the hands of just a few global incumbent mobile companies. This has concrete consequences in terms of the diversity in the ecosystem, giving fewer opportunities to alternative service providers. Community-led non profit providers, for example, can offer connectivity suited for the needs of local groups, including access to networks that are not connected to the global Internet or non-Internet information networks like community radio (Baca-Feldman et al. 2017, Morales and Figueiredo, 2020, Prudencio, K and Bloom, P. 2021). If communities do not have sufficient access to unlicensed and open spectrum or if most of the spectrum is licensed to global private companies, none of this can be achieved.
Framing Internet connectivity as a purely development issue is problematic because it ignores that minority communities may not be part of the formal economy – actually, in some contexts, indigenous populations and ethnic minorities tend to be in dispute with local industrialization programs that take their land and destroy their communities and environment. The economic development focus might also ignore the intrinsic political issues behind inequalities that the digital divide just increases, since it considers the Internet as a mere product available for those that can pay, ignoring that access to information, participation in public life and free expression online should be a right to all. It also provides a limited understanding of what the Internet can offer and limits it to a few global services and one size fits all solutions (Hatmaker, 2015).
The Internet of the rich and the Internet of the poor: zero rating and the implications to the right of access to information
The most popular zero-rated platform in Brazil is the communications application WhatsApp, used by 9 in every 10 Brazilians who connect to the Internet (Fenelon & Torresan, 2018). WhatsApp is not subject to media regulation and journalism standards. This means that its parent company, Meta, has no obligation to implement fact-checking measures or quality control standards related to the content that circulates on its platform (Medeiros & Singh, 2020). This means that if a user wishes to cross-check or verify the information received on WhatsApp to ensure its accuracy, they must be willing to invest part of their data allowance to search on a platform that may not be zero-rated.
Different scholars show how these platforms have been exploited by political actors in ways that interfere in democratic processes in Brazil (Anita Baptista et al., 2020; Panho et al., 2021). Political campaigners and disinformation perpetrators make broad use of zero-rated platforms to create echo chambers to avoid fact-checking in due time, making these tools a threat to open political discussion in the country.
Moreover, the dependency on zero-rated services contributes to deepening the right to access information inequality between people of different income backgrounds. When verifying information outside of zero-rated platforms becomes impossible for those who cannot afford it, access to an open Internet and the capacity to critically analyze and consume information becomes a privilege of those who do not need to be reliant on these types of plans. This environment creates an Internet that discriminates between users, where only a select few can afford open and meaningful access to content and can fully exercise their right to access information. Meanwhile, zero-rated echo chambers present the perfect environment for mass disinformation campaigns. The way that Whatsapp has been weaponized by disinformation perpetrators in Brazil and other countries in the region during election periods in recent years is a consequence of the unequal access that started at the infrastructure level.
Misinformation as the outcome of a privatized Internet infrastructure
The widespread practice of zero-rating is impeding infrastructure expansion exactly in the areas where the lack of Internet connectivity is still a very present issue. Promises made by governments and the private sector under the discourse of socioeconomic development are not necessarily followed by real action.
Internet access does not directly correlate to connectivity; even when the infrastructure technically exists, other important factors can interfere with the meaningfulness of the connection, such as affordability, digital literacy, relevance of available content, etc. Prepaid plans that have low data caps but offer zero-rated applications do not provide meaningful connectivity. It is convenient for mobile network operators and governments to include people that are connected through these types of plans in the statistics that are compiled to indicate the rate of connectivity expansion, but doing so would paint an inaccurate picture of the current situation. Up-to-date, publicly available, reliable official information that disaggregates these statistics based on the types of service contracts that are issued (i.e. prepaid vs. postpaid) and the incidence of zero-rating practices is necessary, especially in regions where the lack of infrastructure is felt more intensively, so that the state of connectivity can be more accurately tracked and studied.
The greater reliance on mobile network operators and other large incumbents to provide connectivity harms people and communities. Connectivity measures that expand the powers and resources of mobile network operators and other large incumbents have direct and adverse impacts on the right to freedom of expression and access to information.
It is important to understand the social and political implications of Internet infrastructure service provision and the current scenario of its extreme privatization and market concentration, since it is now more than ever a prerequisite for exercising personal and collective agency – a condition that will only become more important in the face of the current and possible future pandemics.
 Just recently we have seen a positive outcome of the coordinated work of local civil society working together with the National Telecommunications Agency in Brazil (ANATEL). It was focused on encouraging community networks, supporting small providers, and conducting research aimed at the population, particularly in rural, isolated areas with difficult access and of small municipalities to map the demand and gather inputs to contribute to regulation that meet the local needs https://www.apc.org/en/pubs/policy-brief-and-recommendations-enabling-environment-community-networks-brazil . Other case studies on initatives in the region can be found at https://www.internetsociety.org/es/resources/doc/2018/redes-comunitarias-en-america-latina/
 The fight for a spectrum open to all started more visibly with the pirate radio activism from the 60´s and 70´s. With the Internet connection and media convergence to digital, it was updated to other kinds of communications (including radio). Some seminal statements can be found here https://web.archive.org/web/20080621000315/http://www.acmqueue.org/modules.php?name=Content&pa=showpage&pid=37 and http://www.aaronsw.com/weblog/000737
 Some of the conflicts between local communities and global infrastructure projects can be found in the Environmental Justice Atlas https://www.ejatlas.org/ . More recent attacks against indigenous communities and the current federal government in Brazil are related in Mark Harris for Open Democracy https://www.opendemocracy.net/en/democraciaabierta/the-war-on-indigenous-rights-in-brazil-is-intensifying/
For a customized map of areas where zero rating practice is available, see https://public.tableau.com/app/profile/zeroratingcts/viz/zeroratinginfo/Painel1 as part of the http://zerorating.info/ , an outcome of the dynamic coalition created in 2019 during the Internet Governance Forum of the UN.
Anita Baptista, E., Rossini, P., Veiga de Oliveira, V., & Stromer-Galley, J. (2019). A circulação da (des)informação política no WhatsApp e no Facebook. Lumina, 13(3), 29–46. https://doi.org/10.34019/1981-4070.2019.v13.28667
Baca-Feldman, C. et al. (2017) El Espectro radioeléctrico como bien común: una reflexión en torno a la comunalidad y las redes celulares comunitarias en Oaxaca, México. ALAIC, v.14 n.26 http://revista.pubalaic.org/index.php/alaic/article/view/907
Basri, N. (2019) Zero rating: not the best, but the best of the worst. (2019) Journal of Law and Social Change, 26th November 2019 https://www.law.upenn.edu/live/news/9597-zero-rating-not-the-best-but-the-best-of-the-worst
Fenelon, P., & Torresan, V. (2018, December 1). The Rise of WhatsApp in Brazil from Messaging App to Sales Tool. LABS English. https://labsnews.com/en/articles/ecommerce/whatsapp-in-brazil/
Instituto Locomotiva & Instituto Brasileiro de Defesa do Consumidor. (2021). Barreiras e limitações no acesso à Internet e hábitos de uso e navegação na rede nas classes C, D e E. https://idec.org.br/sites/default/files/versao_revisada_pesquisa_locomotiva.pdf
GSMA (2022) Spectrum Policy. GSMA https://www.gsma.com/spectrum/policy/
Hatmaker, T. (2015) Facebook is becoming synonymous with ‘Internet’ to global users. Dailydot, 9th February 2015 https://www.dailydot.com/debug/facebook-internet-perception-global/
ITU Facts and Figures 2021. https://www.itu.int/en/ITU-D/Statistics/Pages/facts/default.aspx
Knight, P. T., Feferman, F., & Foditsch, N. (Eds.). (2016). Banda larga no Brasil: passado, presente e futuro. Figurati.
Lefevre, F. (2017). Zero-rating, a internet dos pobres https://flavialefevre.com.br/pt/blog/zero-rating-a-internet-dos-pobres
Morales, E. and Figueiredo, G., ed. (2020) Tejiendo desde la contrahegemonía. Medios, redes y tic en América Latina. Universidad Nacional Autónoma de México, Instituto de Investigaciones Sociales http://ru.iis.sociales.unam.mx/bitstream/IIS/5818/2/Tejiendo_contrahegemonia.pdf
OECD. (2018). Bridging the rural digital divide. OECD. https://doi.org/10.1787/852bd3b9-en
Panho, I. A., Bastos, M. D., & Silva, G. F. (2020). O uso do WhatsApp nas eleições de 2018 e as lacunas teóricas da Justiça Eleitoral. Tríade: Comunicação, Cultura e Mídia, 8(18), 173–197. https://doi.org/10.22484/2318-5694.2020v8n18p173-197
Prudencio, K and Bloom, P. (2021) Mantenlo análogo: parámetros para una exclusión voluntaria de la Ia conectividad. Rhizomatica, 9 June 2021 https://www.rhizomatica.org/mantenlo-analogo-parametros-para-una-exclusion-voluntaria-de-la-conectividad/
Recuero, R., Soares, F., & Vinhas, O. (2021). Discursive strategies for disinformation on WhatsApp and Twitter during the 2018 Brazilian presidential election. First Monday. https://doi.org/10.5210/fm.v26i1.10551
Regional Center for Studies on the Development of the Information Society (Cetic.br). (2021). ICT Households Survey 2020: Executive Summary (ICT Households). https://cetic.br/media/docs/publicacoes/2/20211124201635/executive_summary_ict_households_2020.pdf
Star, S. L. (1999). The Ethnography of Infrastructure. American Behavioral Scientist, 43(3), 377–391. https://doi.org/10.1177/00027649921955326
Star, S. L., & Ruhleder, K. (1996). Steps Toward an Ecology of Infrastructure: Design and Access for Large Information Spaces. Information Systems Research, 7(1), 111–134. doi:10.1287/isre.7.1.111
Santos, B. (2018) Eleicoes 2018 no Brasil, entre a polarizacao política e a fraca protecao dos dados dos cidadãos. B https://medium.com/codingrights/eleições-2018-no-brasil-entre-a-polarização-política-e-a-fraca-proteção-dos-dados-dos-cidadãos-c1e51051f1e9
Santos, W. et al (2015) Programas de gratuidade no acesso à Internet: cotrovérsias e indefinicoes https://www.researchgate.net/publication/322600630_Programas_de_gratuidade_no_acesso_a_Internet_conceitos_controversias_e_indefinicoes Song et al. (2019) Innovations in Spectrum Management. Enabling community networks and small operators to connect the unconnected ISOC, https://www.internetsociety.org/wp-content/uploads/2019/03/InnovationsinSpectrumManagement_March2019-EN-1.pdf