Introduction

This article proposes an analytical journey on the Latin American Internet Governance Forums (LACIGF) between 2018 and 2021 paying special attention to the role of Internet intermediaries in content moderation for the exercise of human rights, in particular, of freedom of expression. The objective is to study the approaches of the actors who participate on behalf of the State, the market and civil society.

This is a qualitative and exploratory study from the perspective of public communication policies which allows us to investigate the power relations between the State, civil society and companies in the control and operation of the media system and communication and information platforms. Likewise, the approach to human rights is resumed, in particular, freedom of expression.

This work is organized in four sections. Following this introduction, the methodological strategy is presented. The theoretical framework is shown below and the third section addresses the analysis of the regional forums in the period 2018-2021. Firstly, an overview of the different topics that have risen in the four years studied is presented. Out of all, we have focused on the study of the sessions that addressed human rights issues on the Internet in order to account for the actors and their stance on this aspect. We argue that the problem of private content moderation and its regulation appears incipiently on the agenda of the LACIGF between 2018 and 2019, but in the following two years the approach to human rights moves from freedom of expression to privacy. Part of the explanation for this may be the lack of consensus among stakeholders on how to approach the circulation of speech online.

For the discussion on the forums, it is also proposed a reading that links them with global formulations based on the review of bibliography on the topic. Finally, the conclusions of the work are presented.

Methodological strategy

The methodological strategy consisted of reviewing the videos and/or reports of the ^{LACIGF (LACIGF 2018}, ²⁰¹⁹, ²⁰²⁰, ²⁰²¹⁾ from the selected years in order to investigate topics, actors and positionings. The forums, held in the 2018-2021 period, had a total of 34 sessions, plus the opening and closing talks of each event. Of all of them, specifically four videos -and their corresponding shorthand versions- were analyzed. They referred to the sessions dedicated to human rights. In this regard, it is stated in the work that debates on human rights on the Internet are more incipient on the agenda of regional forums. As can be seen in the analysis below, all the topics addressed in the events and the specific analysis of the sessions that are part of the object of study of this article are summarized in Table 1.

For the study, a survey of the agenda of the event’s sessions in the four selected years was first carried out in order to identify all the topics addressed, their continuities and displacements. Of all the topics, the panels that explicitly referred to human rights and freedom of expression in their title were then selected to carry out a more in-depth analysis of the subjects presented in those sessions, participating actors and their positioning. The reason why these sessions were chosen for analysis was because the problem of private content moderation initially appeared in those panels.

It is worth clarifying that for the analysis of themes the study period was 2018-2021, although for the investigation of actors and their positioning in human rights panels the scope was 2018-2020. This is because those years are completely recorded on the official LACIGF site, i.e., the videos of the sessions and the reports of each one. On the other hand, for the 2021 forum, the agenda is incomplete since not all the panelists have been identified, which is also not made explicit in the reports of the event. At the same time, audiovisual recordings of these sessions are not available. However, this difference in the corpus of topics and actors does not significantly affect the results of the work.

The survey and analysis of the Forums is complemented by the review of the literature and existing documents on the subject. This was done with the objective of establishing a relationship between the regional approach and global debates on the regulation of Internet intermediaries and content moderation.

Internet governance and discussion models

The concept of governance refers to the set of formal and informal institutions, rules, agreements and practices of state and non-state actors, as well as the decisions and behaviors that have a constitutive effect on society (^{Braman, 2004}). Although the problem of global communication governance has a history in media policies, it gained notoriety with the development of the Internet.

^{Flew (2021)} points out that the most important distinction for Internet policy is not between regulation and governance but between provisions that are enforced by external agencies and have some type of negative sanction regarding non-compliance with laws, and regulations. These provisions operate largely on implicit understandings about appropriate platform conduct and the promise of better corporate behavior. This distinction is fundamental when evaluating civil society intervention in Internet governance.

Civil society is heterogeneous, diverse and represents independent and supportive social participation, faced by institutions oriented by instrumental logic, whether of State domination or the search for market profit (^{Cohen & Arato, 1992}). It functions as a space for associations and citizen action that seeks to defend, protect and/or expand rights (^{Sorj, 2012}). In this work, we will limit our research to the representation of civil society that actively participates in the LACIGF forums and is mainly composed of organizations linked to the defense of human rights in the region and business chambers interested in the topic.

The final document of the World Summit on the Information Society (WSIS) emphasized the participation of multiple stakeholders in decision-making on the evolution of the Internet (governments, private sector and civil society), not only taking into account technical but also political aspects (^{ITU, 2005}). This marks a turning point compared to previous modes of discussion that contemplated Internet governance focused on technical aspects. It also included views from the global south beyond the initial presence of the central nations that were involved in technology developments.

In the discussion process evolution, it is possible to identify three models of Internet governance. The first one is the “consensus model”; in the early years, this was typically used by technical parties to reach agreements on the pillars of technological architecture. Then there is the “US-led multi-stakeholder model” which was developed around the Internet Corporation for Assigned Names and Numbers (ICANN) and institutions that promoted the development of the World Society Summit Information Technology (WSIS) such as the International Telecommunication Union (ITU) and the United Nations Educational, Scientific and Cultural Organization (UNESCO), which included the central role of the North American government over the powers of ICANN and the expansion of topics addressed that go beyond infrastructure management to also address new topics such as online human rights and access.

Finally, it is noted that as from 2014, within the framework of NETmundial and the transition of the Internet Assigned Numbers Authority (IANA), a new “mixed global” model emerges, which accounts for the combination of multiple stakeholders and spaces of discussion of multilateral policies, as well as mechanisms that are developed at the national level to interact at the international level (^{Aguerre & Galperín, 2015}; ^{Aguerre et al., 2018}).

As ^{Cortés (2014)} points out, Internet governance must be understood as a game of power balances and a social phenomenon, which allows us to move away from a merely institutional vision, without implying discarding it:

The backbone of Internet governance is its technological configuration, not only because it shapes the digital environment and conditions the user’s behavior, but also because it determines the power of the different actors. It is a symbiotic relationship: technological configuration distributes power and power determines technological configurations (p. 4).

On the other hand, we have warned that the internationalization of debates on Internet governance and the strengthening of technical and economic perspectives has contributed to the weakening of social participation in the debates (^{Mastrini et al., 2013}). In turn, ^{Segura and Linares (2022)} state that there are important inequalities in terms of political and economic power on the part of the different actors and that international participation becomes extremely difficult for many national and local civil society organizations. Hence, the authors distinguish the multi-stakeholder model in the discussion from the participatory model in policy making. In fact, the materiality of the governance process itself favors the existence of over and under-represented sectors.

We turn again to ^{Cortés (2014)} and agree when he points out that:

Regarding governance, there are diametrically opposed views. While some actors see the Internet as a public space where the general interest, human rights and democratic debate should be promoted, others understand it as a space for innovation and commercial exchange, subject to the laws of the market and free competition. Of course, in between these positions there are all kinds of approaches (p. 5).

The WSIS stimulated the participation of multiple actors interested in Internet governance and for its continuity proposed the creation of the Global Internet Forum (IGF) as a permanent discussion space. Since 2006, 16 annual forums have been held in which interested actors addressed critical Internet management issues.

In addition to the global debate forum, the Latin American Internet Governance Forum (LACIGF) was introduced at the regional level, the first edition was in 2008 and the forum has been held annually since then. In addition to the global forum and the regional ones, there are forums at the national level. It is possible to identify a first wave of national IGFs in 2014, after the Snowden statements, and a second wave in 2016-2017 (^{Aguerre et al., 2018}).

The LACIGF were organized with a multistakeholder criterion, with an organizing committee in which three representatives from the technical area, three from the countries of the region, three from civil society, and three from companies participated. The topics proposed in each meeting arise both from the initiative of the organizers as well as from emerging topics resulting from an open consultation with interested people and institutions. On some occasions, themes developed in the global IGF have been followed.

The financing of the meetings is provided mainly by contributions from the Brazilian Internet Steering Committee (CGI) and Google, to which are added contributions from the Internet Society (ISOC), the Latin American and Caribbean Internet Address Registry (LACNIC), and local organizers.

Although it is difficult to measure LACIGF impact on the implementation of regional policies, it is clear that they constitute a beacon of reference on the main discussions on Internet governance.

Digital rights and freedom of expression

The Universal Declaration of Human Rights (UN, 1948) established the fundamental human rights that must be protected throughout the world, including civil, political, social, economic and cultural rights. In our area, it is worth highlighting the right to freedom of information and opinion, privacy, participation in cultural and artistic life, among others.

Digital rights imply the protection and realization of existing rights already enshrined in the universal system of Human Rights and in the inter-American system (which applies to Argentina) in the context of new digital and connectivity technologies. These are the rights to access, use, create and publish through digital media, and also the rights to access and use electronic devices and telecommunications networks (^{Bizberge & Segura, 2020}).

International organizations such as ^{UNESCO (2015)}, the United Nations Rapporteurs on Freedom of Expression and the Inter-American Commission on Human Rights (IACHR) of the Organization of American States (OAS), have affirmed in joint and individual statements that the rights that apply offline are valid online (^{OAS, 2011}, ²⁰¹⁷).

In particular, the rights to freedom of expression and access to information on the Internet resume to the tradition of the human right to information based on the principles of access, plurality, diversity, participation, and equity (MacBride, 1980). It is especially necessary to protect access and use of digital technologies and connectivity, as well as the production, access and management of private and public data (^{Bizberge & Segura, 2020}). In this work we try to analyze how these issues have been treated within the framework of the meetings organized by the LACIGF.

LACIGF analysis

The missing themes. What is not talked about in institutional spaces

Although the list of topics addressed in the LACIGF is extensive, we can find a set of problems that are partially noticed, or that are not addressed at all. Among the first, we refer to those issues that are addressed in the LACIGF but leave aside some conflictive aspects. In the second group are those topics that have no place on the agenda yet.

As has been noted, the problem of freedom of expression has frequently been considered within the LACIGF. However, the approach to this issue has not brought into discussion the concept of freedom of expression enshrined in the first amendment of the United States Constitution, which considers that only the State is responsible for exercising censorship. Given that the main content platforms are governed by this concept, some authors such as ^{Wu (2018)} argue that the first amendment has become irrelevant for the coercive control of political speech and propose protecting discussion channels in the Internet era.

For its part, the analysis of inequalities and gaps is generally approached as a technical problem, which depends on investment and does not question the role that the countries of the region should assume in the development of the Internet. This aspect is repeated in the treatment of the digital economy and innovation. As we have already pointed out (^{Mastrini, 2010}), there is little discussion about the need to think about the development of the Internet with a perspective that includes the sense of reducing the digital divide and the need to increase independent production capacities.

As mentioned above, the problem of private content moderation by actors with great market power and its ability to affect the democratic debate lost its incipient institutionalized place on the forum agendas. This does not mean that the issue is no longer part of the public discussion, the concerns of civil society or the focus of attention of legislators and governments, but it does mark a displacement, or lack of consensus on its inclusion in the institutional agenda, of debate at the regional level that later leverages its positioning on a global scale. The issue, already controversial itself, has attempted to be positioned by some civil society actors, but it has not yet gained “institutionality” in the forum. Other topics do not appear altogether in the programs of the regional Internet governance forums in Latin America.

An aspect that is not addressed, although it is evident to everyone, is the growing ownership concentration in some Internet resources. For example, the market concentration of platforms and the way in which it affects democratic functioning and/or competition have not been the subject of specific treatment in these forums although it is part of the proposals that cross the different axes mentioned. Although the power exercised by certain platforms is recognized, there is no consensus on how to approach it (we will return to this below). In this regard, as it has already been pointed out by ^{Srnicek (2018)}, the most important platforms are building infrastructure and spending money to buy other companies and invest in their own capabilities. They are becoming owners of society’s information, so the monopolistic tendencies of the platforms should be taken into account in the analysis of their effects on society in general.

Outside the LACIGF agenda, the problem of concentration has found a partial echo in the proposals of the Freedom of Expression rapporteurs, who more or less recently have recognized in their Joint Declaration that private control of the Internet is one of the three challenges for the next decade as a threat to freedom of expression (^{OAS, 2019}). The issue has remained mainly in the hands of some civil society organizations and academic sectors (^{Maurício et al., 2021}), which have not had the strength to include it in plenary debates.

A less problematic perspective such as the promotion of competition has also not found an echo in the LACIGF. It can be expected that based on the European regulation on digital markets these principles can be discussed in future forums.

Finally, the way in which Internet development affects the media ecosystem, weakening traditional news producers has not been addressed either. In this regard, news production could be positioned as a non-rivalrous and non-exclusive public good. As ^{Picard and Pickard (2017)} state, the social production of information constitutes a public good with positive externalities that transcend the economic aspect. In this regard, public support for the production of local and regional content, as well as the production of high-quality news and entertainment, is identified as another of the issues absent in the agenda.

Actors and their positioning

After having reviewed the topics, we will focus specifically on the analysis of actors and positioning in the sessions dedicated to “freedom of expression”/“human rights” in the period 2018-2020. For this purpose, a panel on this axis was identified each year. The analysis of actors and their positioning was carried out by watching the videos and/or reading the event reports on the said sessions.

As mentioned in the methodological section, the period included in this case is 2018-2020 because the complete records are available for those years. On the other hand, in 2021 the videos are not available and the reports are incomplete to be able to identify the participating actors. The presentation of actors and positioning in Table 2 has been done taking into account the four large sectors represented at the event: State, Companies, Civil Society and Technical Community.

The approach to private content moderation and the need or not for state regulation varies both in relation to the actors analyzed as well as in relation to the type of speech it refers to.

In 2018, LACIGF was marked by the general elections in Brazil; in this regard, the problem of freedom of expression and content moderation focused on disinformation campaigns and their impact on democracy. Although it is left out of this article analysis, it is worth mentioning that the topic was not only the subject of the session but also the special panel entitled “Polarized pluralism on the Internet and the risks for Latin American democracies”, whose speaker was the then UNESCO Regional Office for Communication and Information Councilor, Guilherme Canela.

From the survey on the participation of the different actors in the session on freedom of expression, a homogeneous position is noted between representatives of the State, companies and civil society who privilege an approach based on digital literacy in electoral contexts.

The positioning of Facebook is worth noting, which in the face of an “accusation” of exercising censorship (removal of speeches or profiles), ends up making its editorial role explicit:

Facebook does not remove or delete profiles... We reduce the reach of that [false] news on the platform, we do not remove it, but we reduce its reach, the user receives the notification that the information was reported as false and the user decides if they want it shared or not (Mónica Guise Rosina, Public Policy Manager of Facebook Brazil) (^{LACNIC RIR, August 1, 2018}).

This statement fully accounts for the role of the platform (regardless of the content in question) to make expressions visible or invisible.

The alignment of some civil society organizations participating in the debates and companies (through the chamber that brings the platforms together) that describe state intervention as “dangerous” and warn that “over-regulating” could lead to state censorship. This view underlies that for the majority of LACIGF participants, state censorship presents more risks than private censorship or moderation whose rules are opaque to the entire population.

Unlike the 2018 edition, the 2019 forum also addressed the problem of private content moderation in a more general way as well as the debate on state regulation/self-regulation of platforms.

From the actors’ presentations a “groping” approach emerges, with more questions than certainties. Regarding government interventions, part of the business sector and technical community began to mention the need for state intervention, although without a clear purposeful agenda of how to do it. Paradoxically, this is especially clear in the intervention of representatives of the state sector. The common element between the different actors -in particular, companies and the technical community- is the position on non-prior censorship. The technical community proposes the combination of state regulation and literacy and warns about the risks of criminalization of online expressions. The latter must be understood in the framework of the Brazilian context in which various bills were being discussed in Congress proposing the criminalization of users and platforms involved in spreading disinformation, for example, through the creation of a new penalization type for those who share “fake news” or the immediate removal of content that refers to certain topics and that prohibits encryption in messaging (^{Intervozes, 2019}).

For its part, the role of civil society has been to raise provocative questions to promote debate and get out of the dichotomy of excessive moderation-no private moderation. The questions regarding the public interest and commercial value of certain content aimed to realize that some levels of moderation are desirable but that the solution on how to address the problem cannot come exclusively from organizations such as Facebook’s Supervisory Board.

Faced with provocative questions from civil society and the government’s position of uncertainty about how to intervene, Facebook’s presentation aimed to validate its own operating mechanisms by invoking the participation of multiple stakeholders in the development of community standards and the role of the Advisory Council. In the words of María Cristina Capelo, Director of Safety and Wellbeing for Facebook in Latin America: “The Facebook Advisory Council is a step forward towards transparency and accountability because the decisions made there will be binding for the platform and that implies a process of reviewing internal policies”, Thus, the participation of experts and other actors in the analysis of cases but also when developing its internal rules is part of the company’s legitimizing discourse in making decisions about speech moderation.

As regards the LACIGF 2020 edition, as mentioned above, the axis of the debate changes completely: from the problem of speech moderation, impact on freedom of expression and whether or not state regulation is appropriate, centrality is given to privacy, especially violated in the context of the pandemic.

From the government sector, the interventions focused on cases regarding the way data was collected and processed, either for the allocation of social assistance (Panama) or difficulties in implementing electronic voting in a context in which the isolation measures made it impossible to attend the elections (Ecuador).

For their part, from international organizations (RELE-CIDH) and civil society, an alignment was evident on concerns regarding the collection of population data by governments in a context of health emergency.

In this regard, it is possible to affirm not only the displacement of the right to freedom of expression (circulation of speech online) compared to the right to privacy in debates, but also a positioning of the State, international organizations and civil society that puts the focus on the dangers of state intervention, but the role played by the private sector, in particular, large platforms, in guaranteeing privacy in the context of the pandemic and beyond, has not been problematized.

The interventions of the private sector were embodied by platforms (Facebook) and “telcos” (through the ASIET chamber). While the former presented corporate initiatives on privacy, the latter emphasized the role of digitization and higher levels of Internet access for economic growth (and social well-being as a consequence) but without specifically alluding to the privacy problem.

Finally, we must mention the absence of academia among the speakers of these debates on human rights (or in some cases subsumed within the space of civil society, for example, CELE with its double role of being a Center within a University) and the limited participation of the technical community.

In conclusion, based on what has been stated about topics, actors and positions, it is possible to affirm that the problem of private content moderation and its regulation appears timidly between 2018 and 2019, but in the following two years the problem shifts from freedom of expression towards a broader approach to human rights, although the focus is on privacy.

This could be explained by the lack of consensus among the different stakeholders on the approach (state regulation/self-regulation) of content moderation rules by large platforms. While in the case of disinformation campaigns in electoral contexts, a shared approach to education (non-regulatory solutions) is privileged, when content moderation is considered in a broader sense, differences and questions arise.

On the other hand, problems such as data processing and privacy -especially in the context of a pandemic- find a higher level of consensus and shift the axis of the debate in a double way. In addition to what we point out on the subject, the displacement is also observed in terms of perspective: what is problematized here is the role of the State, but the performance of the private sector is not questioned either in the context of the pandemic or beyond exceptionality.

Finally, there is a shift in the debate regarding the nature of state intervention. Although at first some positions support the capacity of states to act as guarantors of the social circulation of online discourse, in most recent meetings positions that describe state intervention as a negative limit to privacy and individual freedom have strengthened, in line with the United States’ characterization of freedom of expression.

Final words

This work analyzed the themes, actors and their positions in the LACIGF that took place in recent years, paying special attention to the approach to human rights. We understand that it is an important issue to understand the Internet governance process in the region. The article then reflects the debates institutionalized in the Forums, although in no way exhausts the study of all the positions of all those interested in the topic; not even from governments, companies and civil society organizations, but represents those who actively participate in the forums.

The study of actors and positions reflects the possibilities of a multi-stakeholder organization where issues are agreed upon by its organizing committee. Its agenda is the result of negotiations between the parties, although it permeates proposals arising from consultative forums and other participation spaces.

In general, the topics addressed replicate the themes addressed in global forums, although they may give it a regional nuance. In this regard, the analysis of the topics presented shows a balance of interests and at the same time a limited freshness and controversy in the approach. By prioritizing the logic of consensus, differences are usually mitigated. This can be seen as one of the limitations of multistakeholder governance form of discussion.

Regarding the topics discussed in the forums, one of the main conclusions is that human rights entered the debate in the last five years, but they still represent a small space on the topics of the agenda. The greatest thematic continuity was found in aspects such as access, the digital economy and the institutional design of the LACIGF forums. The second aspect is that the initial concern about content moderation by large platforms was soon displaced by concerns about privacy in a pandemic context.

This movement can be explained by the dynamics of the Forum itself, which follows the multi-stakeholder governance model, in which, as we said, the agenda is defined by consensus based on the topics on which there are greatest points of agreement and where the differences are reduced.

Although certain nuances could be observed in the positions of each type of actor (State, Companies and Civil Society) and even within each subgroup, in the case of the civil society that participated in the human rights sessions, it has been noted a positioning that regards state intervention (both in relation to freedom of expression and privacy) with suspicion.

If, as proposed in the theoretical framework, civil society organizations are distinguished from market and State actors in the search for rights claims, what was observed in the analysis is that civil society organizations that participated in the tables dedicated to human rights were closer to those of companies in these institutionalized spaces. This could account for an agenda co-opted by corporate interests and a freedom of expression agenda that focuses almost exclusively on the risks of the state sector, with few questions about the role of private companies in the violation of rights.

Faced with these approaches, the positioning of States and international organizations such as the Rapporteur on Freedom of Expression have been more “shy.” Its presentations in the different years were limited either to the recount of measures adopted on the topics raised in the session or to point out that their intervention is not necessary (for example, in disinformation campaigns) or exposing general ideas about the need to regulate without clear answers about the scope.

Another element that emerged from the analysis has to do with the identification of the specific actors who participated in the panels. Although the representation is diverse, it can be seen that there are members who are active in all sessions, while others are so occasionally. It was evident that on the part of social organizations and States there are institutions that vary year after year. On the other hand, in the case of the private sector, the sustained participation of Facebook was observed during the three years (in addition to the alternation of the business chambers of telecommunications and Internet organizations). On the one hand, their public participation and knowing first-hand their actions and positions on the issues is extremely important; but, at the same time, that same participation can be read as another lobbying and impact strategy on the public debate. Finally, the scarce presence of the academic sector and the technical community in debates on human rights in the period studied is highlighted.

For companies, participating in forums serves to legitimize their points of view and support their positions, in most cases contrary to any state intervention. Civil society contains a dispersed representation in which each organization tries to influence the debates with the topics that interest each of them the most. Although there are informal ties of debate within them, not all organizations that participate in the LACIGF do so in a coordinated manner. Finally, governments constitute the weak link in participation in LACIGF forums, since their representation is changing and, in most cases (with exceptions, of course), they follow the events. We agree with ^{Aguerre et al. (2018}) when they point out that governance practice is not uniform across national initiatives. It remains to analyze in future works how the discussions of the regional forums translate into public policy axes at the national level.

Internet governance is a crucial aspect given its economic, social and political importance, but mainly for its almost permanent presence in people’s daily lives. Understanding the logic of its operation is essential to promote and adopt principles designed to reduce harm and strengthen the activities beneficial to society.