Vitor Rhein Schirato[2]
ABSTRACT
The scope of this essay is to analyze the recent developments of the social media environment and how the State should (or must) act in its control and regulation, given the essential content of the fundamental rights, the means of equilibrium of such rights and the means of action of the Public Administration.
Keywords
social media — fundamental rights — regulation — police power — selfregulation
RESUMO
O objetivo deste estudo é analisar como vem se desenvolvendo o ambiente das redes sociais e como pode (ou deve) o Estado atuar para seu controle e sua regulação, tendo-se em conta o conteúdo essencial dos direitos fundamentais, seus meios de ponderação e os instrumentos de ação da Administração Pública.
Palavras-chave
redes sociais — direitos fundamentais — regulação — poder de polícia — autorregulação
The rise of social media has proven a double-edged sword, sparking as much fear as it does fascination. It is inextricably woven into every aspect of our lives, as we become ever more connected, for better and worse. The uses are endless—from entertainment and keeping in touch with family and friends to catching up on the news and, more contentiously, trading opinions on all manner of issues and events.
A parallel universe has emerged, a virtual space, where users create online personas often far removed from their normal selves. Our lives are increasingly spent online on all manner of activities: from the exchange of goods and services to personal and professional networking and the expression of ideologies and social critiques.
As the virtual world expands, the companies that support it (the big techs) gain ever more economic power and influence. Their importance and revenue increase in proportion to their expanding user base. And this is no accident: social media platforms use algorithms to deliver highly targeted, personalized content that matches supply and demand far more effectively than traditional advertising.
Hence, the crucial element: ever-expanding economic power. In itself, this is not inherently unlawful. However, it opens up increasing space for abusive conduct, necessitating a degree of economic and legal regulation.[3]
However, the economic clout of these companies is not the only issue. Social media companies are increasingly influential in the political area, playing a crucial role in the development of democratic (and anti-democratic) processes.[4]
In empowering billions of people to communicate freely—often recklessly—such platforms shape opinions and even the outcome of elections. Of particular concern is their role in enabling hate speech and expressions of violence, racism, misogyny, and homophobia.
Put simply: communication is much quicker and easier than through traditional channels, accelerating the speed at which opinions and ideas proliferate. Governments and private groups have successfully pushed to limit hate speech in traditional media,[5] but on social networks they are fighting a losing battle. Comments and pronouncements go viral before any action can be taken.
Moreover, algorithms lead to the creation of echo chambers that limit access to other viewpoints and reinforce biases. Debate is stifled and democratic respect for diverse perspectives undermined. Opportunities to consider alternative, potentially challenging, ideas are lost and the space for dialogue diminished as identical or similar viewpoints become entrenched.
In fact, social media use and practices are bound up with the current crisis of democracy, facilitating the spread and consolidation of anti-democratic ideas, including those which threaten electoral systems or target the rights of minorities, which are widely regarded as the cornerstone of any democratic system.[6]
At the same time, there are broader implications: chiefly the protection of freedom of thought and expression, regarded as fundamental to a functioning democracy. Democracy, by definition, runs counter to censorship.
Hence, the issue is thorny: How can we guarantee the effectiveness of the right to freedom of expression and thought while preserving the essential means for the existence of democracy?To what extent should we seek to moderate the expression of opinions without damaging the essential content of fundamental rights?
In response, state regulation is often proposed, i.e., the enactment of norms to govern the use and administration of social media. The justification is the need to impose stringent measures to contain and curtail anti-democratic discourse, hate speech, and the spread of fake news and information.
Prima facie, the idea is compelling, as it is aimed at the protection of key social values and based—at least in theory—on the balancing of conflicting fundamental rights. The problem is not the concept, but rather the means by which it is implemented.
In the context of modern regulatory theory, the applicability of state-mandated restriction and supervision of digital discourse is highly problematic. Traditional police power, for its part, is largely inadequate, with questionable authority and efficacy in cases involving freedom of expression.
Given this introductory context, our objective here is to examine whether state regulation is conducive to social media control. To this end, we will explore the concept of regulation, its objectives and instruments, and assess its applicability to social media usage and practices to determine whether indeed the problem is a lack or a deficiency of regulation.
This is clearly a complex issue that cannot be exhausted in one article, especially as it requires analysis from the perspective of Constitutional, Administrative, Electoral, and Civil Law. In terms of Administrative Law (not disregarding the constitutionalization of this legal area), a number of considerations may be advanced to shed light on the main issues.
To this end, we will address the following factors: (i) the concept of regulation and its purposes; (ii) the conflict between fundamental rights in the context of social networks; (iii) the suitability of regulation, in its most traditional construction, to resolve conflicts between fundamental rights; and (iv) how and by what means the State may act and intervene.
The concept of economic regulation is inconsistent in the legal sphere, especially when considered from a comparative perspective across different jurisdictions. A broader interpretation is feasible, equating regulation with the traditional notion of police power, as is a more limited concept, which restricts regulation to the strictly economic field.
The ambiguity is rooted in the various ways regulation has been understood historically in a range of legal systems. In the US, notably, the idea of regulation emerged as a mechanism of state action intended to organize certain markets, eliminating or correcting structural flaws that prevent their proper functioning (natural monopolies, economic asymmetries, information asymmetries, etc.). In the US, the concept of regulation is intertwined with Administrative Law.[7]
In this context, it represents the main form of state intervention in the sphere of rights and freedoms. Regulation is, par excellence, a state instrument that limits rights in favor of collective goals.
In continental Europe, on the other hand, state intervention in the area of individual rights is much older and more traditional, manifested through administrative acts within the scope of police power. From the earliest days of Administrative Law, a set of restrictions on individual rights has been accepted to ensure the proper functioning of civil society. The idea of police power has become intertwined with the development of Administrative Law itself.[8]
However, police power in this context has a very different dynamic and focus compared to American regulation. It is essentially directed towards the individual, its primary recipient, and stems from a competence established by law, with little (or no) dependence on administrative norms, except where only simple regulation is necessary.[9]
For this reason, Eberhard Schmidt-Assmann argues that European administrative law began at the “retail” level, grounded on the state-citizen relationship, whereas American administrative law took a “wholesale” approach, rooted in the relationship between the state and the market.[10]
In Europe, the idea of regulation only emerged under this name in the 1990s, in the context of state reform, with respect to de-monopolization and privatization. In this framework, regulation serves as a mechanism for fostering market competition in sectors previously dominated by state monopolies, particularly public services and infrastructure.[11]
Thus, we observe a contrast between a broader American-style conception of regulation and a more restrictive European model focused on major economic sectors or areas where collective risks are imminent, such as data protection and public health. In the first case, regulation and police power converge. In the second, there is a clear demarcation: regulation is reserved for the macro-level intervention in sensitive economic or high-risk markets, while police power is limited to the state-individual relationship, restricting and conditioning the exercise of rights.
Viewed from a Latin American perspective—particularly that of Brazil—there is a clear alignment with the continental European system. In this context the tradition of Administrative Law tends more towards "retail" than "wholesale," to use Schmidt-Assmann's metaphor. The notion of police power has evolved historically to designate the circumstances in which the State may limit individual liberties to guarantee collective well-being. The idea of regulation also stems from the market liberalization and privatization of the 1990s.
Hence, this is not merely academic speculation. While police power and regulation may be linked in terms of their substance, as state measures that supervise fundamental rights, their methodologies are distinct in the European tradition, a characteristic mirrored across most of Latin America.
In this model, the exercise of police power by the public administration must have an explicit legislative mandate. Such laws provide the grounds for limiting individual rights in the interest of the common good. Police power depends on the tripartite notion of function-competence-act: 'function' is the legal prescription of the authority to act; 'competence' is the assignment of authority to a specific official; the 'administrative act' is the resulting exercise of power.
In the exercise of police power, there is a greater dependence on parliamentary legality, since the law must create the function and assign the competence, with at least some understanding of the administrative act in question. Obviously, in some cases these boundaries are strictly defined, while in others a degree of discretion is involved.[12] However, it is more strictly bound to parliamentary legality.
In the regulatory field, there is an assumption of greater normative freedom for the Public Administration, depending on the type of activity performed. The function and competence are still assigned by law, but less detail is provided about the actual substance of the administrative act, which is primarily derived from regulations issued by the administration itself, based on its delegated rulemaking authority.[13]
A distinct methodology is at play in regulation, compared to traditional police power. The complexity, technicality, and evolving nature of regulated sectors require more agile and flexible regulatory frameworks capable of adapting constantly to specific goals, such as ensuring market stability or protecting vulnerable interests.
Furthermore, the content of the normative discipline governing the exercise of rights by private individuals differs in regulation from the exercise of police power. In the first case, the focus is much more on defining the conduct expected of regulated entities. In contrast, police power imposes limits and restrictions on the exercise of certain rights, separating lawful from unlawful conduct.
Hence, economic regulation clearly refers to an activity of the state that predetermines how economic actors operate in order to correct certain flaws in strategic markets. It is, undoubtedly, a method of restricting individual rights for the benefit of the community, with distinct processes and procedures compared to traditional police powers.
Evidently, regulation and police power may be classified together, given their underlying nature. Both represent authoritative state actions aimed at controlling fundamental rights.[14] However, they are materially distinct in terms of method and form. Regulation presupposes greater normative power on the part of the Public Administration and predetermines conduct, while police power, based on parliamentary law, establishes limits and restrictions on the exercise of rights.
In this vein, and with regard to the Brazilian context, the following concept of economic regulation applies: State activity is carried out based on normative acts issued by the Public Administration, with broad functions defined by law, which determines ex ante the conduct of officials within specific sectors to ensure their proper functioning, correct their structural flaws, and protect against and mitigate the occurrence of significant risks to the community.[15]
Consequently, regulation in Brazil is strictly focused on specific sectors. The object of regulation, therefore, is private conduct that may compromise the proper functioning of regulated sectors and the protected legal interests at stake.
As discussed, the issues surrounding social networks are complex. Plainly, enormous economic power is involved. Yet the digital market shows no significant failures, as illustrated by the recent rise of Bluesky—a clear example of how new apps can thrive by simply gaining user traction. Competition is healthy. Yet, for all their economic weight, material abuses against consumers or other market participants are relatively rare. The market is clearly alive and functioning.[16]
The dysfunction, which is profound, arises from the non-commercial use of social media—specifically, its role in the expression of thought. The open, often anonymous nature of social media, amplified by the physical isolation of the virtual world, significantly broadens the spectrum of expression. Opinions once kept private for fear of social backlash are expressed with ease online, as the digital environment lacks the immediate social cues that typically enforce normative behavior.[17]
Consequently, established social and legal norms[18] which have evolved to curb hate speech in the public square, are notably lacking in the virtual world. As prohibitive barriers fall away, unrestrained, often controversial, views proliferate. The simple act of posting a comment, directly or using a fake profile, is enough to impact an indeterminately large audience.
Further complicating the situation, the algorithms behind social networks play a huge role in the dissemination and promotion of the most diverse ideas, based on the profiles and habits of the network users. Thus, not only is expression free, but its spread is amplified, reaching untold numbers of people. A racial slur uttered in the physical world has limited reach, impacting only bystanders and, possibly, their immediate circle of contacts. It is restricted in scope. The same comment expressed on a social network, however, has indeterminate, unlimited reach, boosted directly by algorithms in proportion to its degree of success.
This situation presents a clear conflict of fundamental rights. On the one hand, the rights to freedom of expression and thought are enshrined in Article 5, items IV and IX of the Federal Constitution. On the other hand, the rights to gender equality, personal honor, privacy, and religious liberty (Article 5, I, VI, and X), for example, are guaranteed under the law.
Moreover, unfettered dissemination of all forms of expression may engender violations of the Democratic Rule of Law, including persecution of minorities, harassment of opponents, and imposition of anti-democratic regimes. Again, we see a conflict between fundamental rights to freedom of expression and thought and the right to the Rule of Law (Article 1 of the Federal Constitution).
Such conflicts underscore the need for legal interventions tasked with reconciling tensions through carefully established criteria.[19]
Based on accumulated legal experience, at least three potential measures stand out (subject, of course, to modification and refinement): (i) imposing a duty of content moderation upon social media companies through a clear self-regulation mechanism; (ii) avoiding prior restraint and reverting moderation to post-publication legal remedies sought by injured parties; (iii) implementing a priori control mechanisms based on the principle of proportionality.
In the first case, no state measures are employed, premised on the belief that the market is self-correcting. To a degree, this reflects the regulatory framework currently in place.
Under the second model, which, in practice, often overlaps with the first, State intervention is strictly remedial, foregoing prior restraint in favor of two different approaches. The first assumes that the right to freedom of expression cannot be subject to prior restrictions under any circumstances, i.e., intervention must be limited to providing redress for actual and proven harm arising from such expression. To some extent, this is the prevailing mechanism in American law.[20] The second allows for restrictions only after the fact, targeting the withdrawal of content already manifested.
In the third scenario the State may impose measures limiting the rights to freedom of expression where there is a conflict of fundamental rights. These restrictive measures should always be based on proportionality. This is the model adopted in the German system, also widely followed in other Western countries.
A fourth mechanism exists, which is essentially incompatibility with the democratic rule of law, i.e., prohibiting any form of expression on social media in flagrant violation of fundamental rights. This, of course, is a pathological characteristic of dictatorial states, which does not warrant extended analysis.
Hence, the exercise of the right to freedom of expression on social media is highly complex, striking at the heart of the democratic system. But how may we address this complexity and the emerging challenges?
Identifying possible methods does not equate to a clear solution, insofar as a method, in itself, is capable of resolving little; the issue requires the development of instruments of intervention that are both effective and legally acceptable, based on each respective approach. This is the great challenge that lies ahead.
To explore further, we will now analyze each of the three control mechanisms within the Brazilian constitutional context, weighing their relative advantages and disadvantages.
As mentioned, the first mechanism is social media self-regulation. In this system, platforms themselves are responsible for controlling the content and removal of any posts that infringe on the rights of third parties. This is the most common mechanism globally, defended by big techs for almost a decade.[22]
The legal nature of self-regulation is strictly contractual. Users agree to the network terms and conditions when creating their profile, which allows the platform to block, restrict, and remove content deemed offensive to policy rules.
This method, widely supported by the networks, involves an adhesion contract. By agreeing to the terms of service, the user consents that, as a condition of use, some of their posts may be removed in accordance with internal policies. It is a form of prior consent for compliance with the norms imposed by the network.
Thus, the company is responsible for analyzing the content of user posts to verify their compatibility with established policies and, if necessary, impose some form of restriction.
Hence, self-regulation is not strictly tied to or dependent on the legislation of a particular country. In other words, these platforms are not clearly bound by a set of precepts nor by mechanisms guaranteeing specific legal rights.
The values protected by the legal system and those subject to social media self-regulation mechanisms are not necessarily equal or compatible. Clearly, social media policies and legal tenets, such as combating pedophilia and terrorism, will normally align. However, a number of issues that receive specific treatment under national legislation are not addressed by social networks.
Hence, it is perfectly possible for a particular post to be in violation of network policy and be removed, without violating the legal system in force in the country where the user who authored the post is located. The opposite case is equally possible, whereby a post is acceptable on the social media platform but violates the norms of a specific legal system.
Adding to the complexity, posts are not simply random content floating in the digital void. As discussed, algorithms are designed to boost such content, flagrantly amplifying illegal content, to reach an ever-growing audience.[23]
In this light, the usage rules established by the networks are, in effect, a superficial, fragile layer of rights protection. Their effectiveness is dependent on criteria set by the companies themselves, in line with their purely commercial interests. There is no direct alignment between the networks and national legal systems, only circumstantial coincidences.[24]
Hence, while, in theory, self-regulation, based on usage policies established by providers, moderates content and goes some way to combatting various types of abuse, it is a highly unstable mechanism. It depends on what the networks deem acceptable and lacks direct alignment with the particular legal order. Correlation between removable content and the fundamental rights embedded in a system is incidental.
The second method of controlling freedom of expression is exercised ex-post, i.e., after the publication and dissemination of content. This can unfold in different ways, based on different premises, as explained below.
In the first scenario, any ex-post control is compensatory, in the form of reimbursement for direct damage caused to a person or group of people. The underlying premise is that the right to freedom of expression is unfettered, and any form of limitation is inadmissible. If the expression of thought causes proven damage to another person or group of people, however, the right to compensation is available. In summary, a pre-existing lack of constraints on freedom of expression stands in contrast to the conventional rule of civil liability, which imposes consequences for illicit acts resulting from an abuse of rights.
Under this framework, no government body, including the courts, has the power to restrict freedom of speech or the expression of ideas. Any and all forms of expression are admissible, including racist propaganda (such as the spread of Nazi ideology), homophobia, misogyny, and other threats to the rights of an individual or group. However, if such content causes demonstrable harm to any person or group, there is a clear right to compensation, based on the general rules of civil liability.
The prevailing understanding is that any and all forms of expression are lawful and should be guaranteed. The only safeguard in this sense is the certainty of the right to compensation for specific and proven damages a third party may suffer. In this light, anti-democratic, racist, misogynistic, homophobic, or other discriminatory content is perfectly acceptable under the umbrella of freedom of expression, provided it does not cause proven harm to a person or group. If it is found to have done so, compensation will be due.
According to a compensatory, ex-post model, a user may lawfully express generalized views against a certain race without targeting specific individuals. However, if this causes definitive harm to a particular person or group, they may sue for compensation.
Broadly, the notion of an abstract conflict between rights arising from the exercise of freedom of expression is rejected. Even if other conflicting rights are guaranteed by the system, legal redress is only available in specific, individualized instances where harm is clearly demonstrated.
This method, again, is prominent, although not uniform, in contemporary American law.[25]
Another form of ex-post intervention may be applied to social media moderation: state action aimed at forcibly removing content that violates other rights. This mechanism is diametrically opposed to the first, since the assumption is the reverse, i.e., that limitations to the right to freedom of expression must be considered.
According to this view, state authorities may order social media companies to remove content due to conflicts between the right to freedom of expression and other fundamental rights implicated in the particular case.
This approach reflects a characteristically German model, rooted in the notion of proportionality;[26] which acknowledges the possibility of conflict between rights, to be resolved in terms of proportionality, i.e., the extent to which one right should be restricted in order to preserve another.
This framework involves a case-specific assessment of freedom of expression in relation to other fundamental rights, determining whether a restriction is permissible in that instance. This analysis considers: (i) whether freedom of expression clashes with other rights; and (ii) if so, whether the proposed restriction meets the criteria of (a) necessity, (b) suitability, and (c) proportionality in the strict sense.
Necessity involves determining the need for a restriction on the right to freedom of expression. Suitability refers to whether the restrictive measure adopted is adequate for its intended purpose and whether another, less restrictive measure may be applied. Lastly, proportionality in the strict sense determines whether the intensity of the measure taken (already established as necessary and appropriate) corresponds to the case in question or whether it is excessive.
Within this framework, the public authority assesses whether a particular instance of free speech has infringed upon the fundamental rights of an individual or collective rights, such as democracy—imposing restrictions ranging from reduced visibility to total removal of the content
This model differs from the others in two key respects: (i) it involves direct application of a legal framework rather than enforcement of policies set by social media providers; and (ii) as a result, the providers are subordinate to the public authority and strictly compelled to remove or alter content as directed
Furthermore, the authority empowered to impose restrictions on the right to freedom of expression may vary according to the jurisdiction under analysis. It may be the administration, as in Germany,[27] or the Judiciary itself, or a hybrid power with jurisdictional and administrative functions, as is the case with the Brazilian electoral justice system.[28]
The central point is the premise that restrictions can be imposed on the right to manifestation of thought (freedom of expression). What is necessary is that a conflict has been established between this right and another, and the restriction imposed is established in line with the principle of proportionality.
The key question is how this mechanism is implemented after particular content has been made public. This is a essentially corrective and, as such, may be combined with remedial actions, as, in many cases, simply removing the content is not enough, as the damage arising from its dissemination has been done. Thus, those harmed by it may be entitled to financial compensation.
Among potential solutions for state oversight of social media, ex-ante control may also be considered—a mechanism characteristic of the police power or state regulation mentioned in Section II.
As outlined, police power imposes limitations or restrictions on the exercise of specific rights or freedoms to ensure social cohesion. In contrast, the concept of regulation entails a state-driven mechanism which pre-conditions the actions of sector agents to ensure full operational functionality[29]
Following this logic, ex-ante mechanisms may function as either state police power or economic regulation, depending on the structure of the state intervention.
Under a standard police power framework, a regulatory instrument (norm or legislation, as the case may be),[30] restricts the manifestation of thought while establishing specific penalties for non-compliance. A generic, abstract restriction on the exercise of the right is provided, with a clear imposition of limits. A public organ is designated as responsible for the verification of compliance with the limits, through oversight instruments.[31] Lastly, if a violation is found, a sanctioning process is implemented, to enforce the respective penalties.
Obviously, any limitation imposed on the exercise of the right to free expression must be supported by the principle of proportionality already discussed. It must be duly balanced and considered. Hence, it is necessary to identify the conflicting rights and the restrictive measures necessary.
This, in theory, is eminently possible. In reality, however, it contravenes the Constitution, insofar as a restrictive norm with such characteristics will likely be configured as a form of censorship. Establishing proportionality in the abstract is, ostensibly, an unattainable goal within a democratic constitutional system. Balancing freedom of expression against other fundamental rights in this way is simply unfeasible, particularly as regards the specific tests of suitability and necessity.
Only in certain circumstances is a priori limitation possible, such as in the case of racist or discriminatory abuse. However, in such scenarios, resorting to police power is unnecessary, as the conduct in question is of such gravity that it is already classified as a crime or serious offense.[32] In other words, in cases where a limitation is clearly proportionate, criminal offenses already exist.
In other cases, limitation ex ante is enormously complex. Article 5 (IX) of the Federal Constitution explicitly states: “The expression of intellectual, artistic, scientific, and communication activities is free, regardless of censorship or license.” In other words, the Constitution forbids any form of censorship, which would occur through state regulatory measures that impose a priori and abstract limitations on the right to the free expression of thought.
This needs clarification. As is widely recognized, censorship entails state interventions designed to block the spread of ideas by individuals or groups. Censorship can be prior, if manifested before the public disclosure[33] or a posteriori, when it retrospectively determines the cessation of something already made public.[34] For our purposes, I will analyze the hypothesis of prior censorship.
In this scenario, a directive identifies which types of discourse should be restricted, as well as explicit criteria. The problem, however, is that this model directly enables censorship, as in a democracy no form of expression is subject to a priori limitation; not even manifestations against democracy itself.
Consider this legal norm: “Any expression of opinion contrary to the democratic regime is prohibited. Violations will result in the immediate removal of any post. Such a rule is, arguably, wholly unconstitutional. Democratic regimes are, in principle, obliged to tolerate criticism. Citizens should even, in theory, be free to express the opinion that the democratic system is ineffective, and an autocratic system should replace it. This should be tolerated and exposed to debate.[35]
Hence, any generic and a priori limitation on the right to freedom of thought and expression, vis-a-vis traditional police power, is, to a greater or lesser extent, unconstitutional, imposing, as it does, prior state censorship.
In terms of regulation, a potential model involves a form of standardization of the exercise of freedom of expression, a specific a priori code of conduct. Again, the main distinction between traditional police power and regulation is that the latter seeks to pre-define conduct to overcome shortcomings in a given sector.
However, there are no clear precedents for social media platform cases. Fundamentally, no ideal standard of private conduct exists to ensure freedom of expression will never infringe upon the fundamental rights of others.
Another example is called for. In capital markets regulation, a core problem is information asymmetry between investment providers and investors. A solution to this is full disclosure, according to which an obligation of prior conduct is established for suppliers to share any and all relevant information with the market. Fulfillment of the obligation of full disclosure, it is understood, will remedy the problem.
In terms of freedom of thought and expression, what conduct should be expected of those involved (including network users and the companies providing the services)? What standards of conduct should be imposed to ensure there are no violations of the rights of third parties? Is a pre-established, definitive framework required? (Clearly, that would amount to a form of censorship). Should vocabulary use or codes of etiquette be imposed? But how to measure the effectiveness of such measures?
Whether they employ traditional police power or modern regulatory tools, ex-ante interventions are clearly fraught with dangers, which may be insurmountable. Indeed, scenarios that require proportional a priori limitation are of such gravity that they already constitute crimes (pedophilia, racism, racial slurs, homophobia, etc.). In less egregious cases, an ex-ante remedy rarely aligns with fundamental rights, since such measures create an expansive gateway for state censorship.
Furthermore, any form of normative action ex ante requires extensive agreement between state officials, social media companies, and users. In this scenario, State action must be taken against both companies and users, imposing a type of joint and several liability. Or, alternatively, a dual level of regulation is imposed, one public, aimed at the network provider, and another private, targeting the companies and the users.[36]
No matter the arrangement, the most complicated issue, from a constitutional standpoint, is establishing clear criteria within a democratic system to limit or predetermine the exercise of the rights to free thought and expression.
Social media activity entails a tripartite legal relationship: of users (citizens), social media platform providers, and the State, the entity responsible for preserving social harmony and sustainability and enforcing legal norms.
In this regard, we observe the legal status of citizens, holders of the fundamental right to freedom of expression; of social media platform providers, which offer services and are entitled to freedom of initiative (article 5, item XIII, of the Federal Constitution); and of the State, which must guarantee the full and complete application of current and effective legal norms.
These somewhat obvious considerations demonstrate the increasing complexity of the situation: those who hold the fundamental right to freedom of expression (citizens) depend on the companies that support social media networks. Therefore, any improper exercise of the right to freedom of expression and thought necessarily involves third parties, who play a key role in defining the irregularity of the exercise of this right and ensuring its continuation.
This raises the question: who is ultimately responsible for damage resulting from the improper exercise of freedom of expression? Is it only the users (who exercise the right) or also the platforms (that make the illegal act possible)?
In the Brazilian context, article 19 of the so-called Marco Civil da Internet (Federal Law No. 12.965, of April 23, 2014) determined that social network support companies are not in any way responsible for the content published by users.[37] For the purposes of Brazilian law, only the latter can be held responsible for the content.
Social companies can only be held liable in the event of non-compliance or improper compliance with a specific judicial order that mandates the removal of content deemed improper in a specific legal proceeding.
The rule in question is of significant importance, reflecting a deliberate choice by the Brazilian legislator to adopt a certain degree of deregulation for social networks. As can clearly be inferred, it is not the responsibility of social media support companies to carry out any form of control over the posts made by their users. Users are free to publish any content they wish. A third-party state body is responsible for making a judgment on its admissibility. If it is deemed inadmissible, the company must simply comply with the decision and remove the published content.
Hence, in reference to the methods discussed, Brazil has adopted an exclusively ex-post control system, with no obligation for the social media companies to act. Regulatory oversight is limited to ex-post intervention by a public authority, which is tasked with assessing the content against the standard of proportionality.
Evidently, however, companies may and do implement self-regulation, based on a service provision contract signed with users, within the framework of subsection III.1 above. However, this procedure is not mandatory and left exclusively to the discretion of each company.
Currently, more than 11 years after the enactment of the Brazilian Internet Bill of Rights and with the rapid development and growth of the market, the aforementioned article 19 is under scrutiny, as, in practice, it leaves the virtual environment excessively open to abuse.
First, it tends to foster a lack of neutrality among social media providers, assigning them an essentially passive role. In this context, a particular network may offer complete freedom for any and all types of expression, creating an echo chamber that would be filtered out under a self-regulatory environment.[38]
Secondly, the Brazilian legislator tends to assume social media support companies play no active role in the spread and propagation of posts. As can be inferred from the aforementioned article 19, these companies simply serve as means of dissemination. They are not responsible for the content published.
Yet, this is patently mistaken. As discussed, algorithms play a decisive role in determining the level of reach of posts. As a result, content does not exist statically on platforms, free from the involvement of the networks that manage it. The companies play an active, vital role in determining dissemination. While posts are exclusive to the author, their reach and scope are bound by the affirmative actions of the social media platform
Thus, companies are not devoid of responsibility. From a legal standpoint, there is clear participation by the networks, who are liable for damage caused to individuals or groups arising from misuse of freedom of thought and expression.
Based on article 927 of the Brazilian Civil Code[39] and article 29 of the Brazilian Penal Code[40], as a general rule, not only the perpetrator, but those who participate in its commission are held criminally and civilly liable. Hence, Article 19 of the Brazilian Internet Bill of Rights is a major innovation, especially considering that the trend in Brazilian law is towards an expansion of civil liability—not reduction.[41]
The matter is so contentious that the Supreme Federal Court (STF) recently addressed the constitutionality of Article 19 of the Internet Bill of Rights (Marco Civil da Internet) in RE 1037396 (Justice Dias Toffoli, Theme 987), a case with general repercussion status.
Ruling on the appeal, the STF, by a vote of 8 to 3, declared the provision unconstitutional. In establishing the legal precedent resulting from the judgment, the Court set forth a series of interim rules for social media providers to remain in effect until the National Congress enacts specific legislation on the matter.
Theme 987 concisely established the following rules: (i) Article 19 of the Marco Civil da Internet cannot be invoked to shield platforms from liability regarding certain types of illegal content and thus providers may be held liable for user posts in these specific cases; (ii) such liability shall exist whenever posts may constitute specific types of crimes enumerated by the Legal Thesis (e.g., crimes against the Democratic State of Law, crimes of racism and discrimination as defined by law, terrorism, incitement to suicide or self-harm, sexual crimes against the vulnerable, and human trafficking); (iii) social media platform providers must maintain self-regulatory mechanisms to ensure the timely removal of content that may constitute crimes; (iv) there is a presumption of liability for social media providers whenever content is remunerated or boosted; and (v) Article 19 remains applicable as written only in cases involving crimes against honor, meaning providers are only required to remove such content following a judicial order. Finally, the STF decision also required the National Congress to legislate on the matter as quickly as necessary.[42]
In Theme 987, therefore, the Supreme Federal Court imposed a system of regulated self-regulation in which social media providers are required to moderate and supervise the content of user posts under penalty of direct liability. If they fail to fulfill their obligations, the Judiciary may intervene, ordering removal of the content and potentially imposing sanctions on the said companies.
What emerges from the conclusion of the STF ruling on Topic 987 is the need for a shift in how social media companies provide their services, as they may now be held liable for participation in illicit acts committed by their users, which constitute an improper exercise of the right to freedom of expression.
To advance in this area, terms of service with clearer rules on the admissibility of content that may constitute crimes or conflict with fundamental rights and a proactive stance in combating them—the removal of posts and even profiles—are required. After all, networks have the most agile and effective mechanisms for identifying potential violations of the law.
Furthermore, the reform of Article 19 of the Brazilian Internet Bill of Rights aligns the policies and terms of use of social media companies with the norms and methods of Brazilian law, resolving, albeit partially, the problems described in subsection III.1 above. The reason for this is simple: these companies will be subject to sanctions in Brazil and, therefore, must have mechanisms in place to ensure their actions comply with the national law.
The previous analysis highlights a contemporary dilemma: social media is an inseparable part of our world yet cannot remain unregulated and free from any form of control. Unfettered and unregulated use of networks undoubtedly leads to functional harm.
The same rationale that led to the establishment of mechanisms for authoritarian intervention by the Public Administration in the 19th century is, arguably, applicable to today, in relation to social media proliferation. Ultimately, the capacity of authorities to diagnose the lawful contours of the exercise of rights and freedoms is grounded in the need to maintain a functioning community.[43]
However, an examination of the tools used to restrict individual rights and liberties developed by Administrative Law over the last two centuries[44]—namely, police power and regulation—shows that none are constitutionally adequate for governing the expression and manifestation of thought in the digital sphere
This happens for a number of reasons, as discussed, summarized below.
In the case of police power, limitations and conditions are imposed on the exercise of a certain right for the greater good. However, the feasibility of pre-defining (a priori) specific restrictions or conditions on the exercise of free speech and the manifestation of thought are deeply contentious. What are the criteria for such restrictions? What is necessary, appropriate, and proportionate?
To date, no obvious criterion compatible with a democratic constitutional system which values basic freedoms has been found. Only dictatorial systems, wherein the right to freedom of expression and assembly is not guaranteed, impose rules that restrict these rights. Democratic societies depend, for their survival, on guaranteeing freedom of expression and the right to express ideas and thoughts. As a general rule, any regulation that limits these rights in a broad and abstract way violates the Constitution.
Furthermore, the case in question is clearly distinct from other forms of administrative supervision, as the ultimate outcome sought through such restrictions remains uncertain. For example, restricting demonstrations of a political nature is clearly unacceptable, just as it is to curb the expression of opinions on everyday societal issues. Such initiatives are ostensibly incompatible with the democratic constitutional system, as the right to express oneself freely, including against the system itself, is foundational.
Put simply, the historical construction of administrative police power identifies clear goals, such as preserving public order, health, and safety. For example, to ensure proper traffic flow on public roads, it is necessary to impose speed limits, prescribe parking locations, establish priorities for pedestrians, etc. Regulations governing street traffic operate with a clear and pre-established objective.
In the case of social media, however, the objective is not clearly known beforehand. Patently, direct threats to the integrity and existence of the democratic system are unacceptable, as is hate speech inciting violence of any kind. But what is the appropriate instrument to control this? Preventing free expression is clearly unconstitutional, as is pre-determining limits on admissible content, given the difficulty of identifying a priori proportionality.
In this regard, although the issue and the desired solution are identifiable (the proper functioning of democracy), determining an effective instrument to ensure this end is complex and challenging. Undoubtedly, the threshold between what is proportional (and constitutional) and disproportional (unconstitutional) is practically impossible to define a priori.
The same applies to regulation. According to the criteria presented, regulation is distinct from police power, as it does not impose limitations and conditions on the exercise of rights; rather, it pre-regulates the expected conduct of economic agents, resolving market failures.
Unlike police power, which imposes clear limits and conditions on the exercise of a right, regulation involves a pre-determination of conduct aimed at preventing behaviors that impede the proper functioning of a given market (for example, the charging of non-discriminatory fees for access to infrastructure).[45]
Yet, much like the challenge of imposing constitutional restrictions via police power, establishing the specific behaviors that must be pre-regulated in order to address the concerns associated with expanding social media use remains a formidable task. As with police powers, the definition of undesirable behaviors to be corrected through regulation is uncertain.
Under current regulatory frameworks, there is clarity regarding the specific behaviors of economic actors that lead to market failures so that the corrective actions required are equally evident. Yet, in the context of social media, there is no clear consensus on which conduct should be restrained or pre-regulated and, more specifically, to what extent. Any pre-regulation of conduct will likely be considered disproportionate and, thus, unconstitutional. Implementation of regulatory action ex ante invariably results in disproportionate restrictions on the rights to free expression and thought.
The only obvious exception pertains to conduct formally codified as criminal. Certain actions and discourse are considered so socially offensive they already constitute a crime, even if they are not directed at a specific person, as with instances of racism, homophobia, and transphobia, among others, which are designated by type.[46]
Consequently, neither absolute deregulation nor stringent ex-ante regulatory models are viable: deregulation will allow grave crimes to be committed without redress (for example, incitement to stage a coup d'état or attacks on a minority). Regulation, on the other hand, presents significant difficulties as regards the substance of ex-ante norms that could dictate how the rights to free speech and manifestation of thought should be exercised. Finding an optimal point that satisfies the requirement of proportionality is seemingly unfeasible.
Given these considerations, the most appropriate mechanism of state action is arguably ex post control, as implemented in a number of countries. In this model, the State, through the Judiciary or a Public Administration body, determines the removal of social media content deemed inappropriate. An ex-post approach transfers to the public official the duty of exercising a proportionality test between the rights to freedom of expression and other fundamental rights that may be in conflict in a specific case.
However, it is simplistic to imagine that only state agencies are able to satisfactorily moderate social media content. The sheer volume of content and the vastness of the virtual environment render this mission almost impossible. Hence the necessity for the involvement of social media companies. Social networks are, and must be, leading players in supervising the virtual environment, undertaking concrete action in the moderation of publications.
Self-regulated mechanisms, therefore, are ostensibly the best means of controlling virtual environments. However, to be effective, regulated self-regulation mechanisms require continuous cooperation between state regulatory bodies and private self-regulatory bodies.
Two elements are fundamental: the establishment of terms and conditions of use by the companies, consistent with the legal systems of the countries in question; and the full submission of these companies to the legal order and system of each country.
Therein lies the greatest challenge of the regulated self-regulationmodel: cooperation of social media companies. Their economic power is, effectively, greater than many countries. At the same time, economic logic determines that their activity is more profitable with fewer regulations. The policies and activities of these companies, furthermore, are increasingly linked to key geopolitical positions and ambitions. Hence the challenge of subjecting them to the authority of a sovereign state.
The national case is paradigmatic in this regard, as Brazil has one of the largest social media bases. Online content plays an increasingly prominent role in the political landscape.
Currently, there is persistent tension between tech platforms and the Judiciary, the latter leading efforts for ex-post oversight in the virtual sphere—in large part because no administrative body has yet been granted the specific mandate to regulate this field. In some cases, there are self-regulating mechanisms in place. In contrast, other platforms (notably, X and Telegram) have no self-regulation mechanism, with restrictive actions exclusively carried out by the Judiciary.
Hence, the ruling on Theme 937, which imposes the duty to consolidate regulated self-regulation of the entire virtual environment, under penalty of financial, even criminal, liability for the social network companies. Whether the rules established by the Supreme Federal Court in Case 937 are a solution to the problem, only time will tell. However, it suggests a more balanced path between guaranteeing the rights of freedom of expression and other, potentially conflicting, fundamental rights, namely, self-regulation regulated through cooperation between the State and private companies.[47]
The aim of this study was to outline the potential features of an effective mechanism for the much-debated regulation of social networks. Evidently, established regulatory frameworks that succeed in other sectors are ineffective when applied to the digital sphere. Hence the need for new methods and instruments. However, these instruments presuppose the cooperation of economically powerful companies with the means and influence to challenge the sovereign authority of states.
As discussed, current mechanisms for shaping the exercise of fundamental freedoms and rights require revision and improvement. Clearly, it will take time to reach an optimal point. A viable path forward has been outlined, but its effectiveness will depend on the will and cooperation of all those concerned.
Lastly, in answer to our original question, the issue is not whether to regulate or not. Some form or regulation is vital, given the proven harmful effects that result from its absence (disinformation, violence, attacks against democratic institutions, etc.). The key challenge is determining the appropriate regulatory methodology, a task that requires further empirical experimentation and observation.
ARAGÃO, Alexandre Santos de. O poder normativo das agências reguladoras. Revista Trimestral de Direito Público, São Paulo, n. 36, out./dez. 2001.
BOLDT, Hans; STOLLEIS, Michael. Geschichte der Polizei in Deutschland.
DANWITZ, Thomas von. Was ist eigentlich Regulierung? Die Öffentliche Verwaltung, Stuttgard, n. 23, 2004.
FUNK, William F.; SEAMON, Richard H. Administrative law. 4. ed. Nova York: Wolters Kluver, 2012.
KINGREEN, Thorsten; POSCHER, Ralf. Grundrechte — Staatsrecht II. 40. ed. Heidelberg: C.F. Müller, 2024.
MARTINI, Mario; MÖSLEIN, Florian; ROSTALSKI, Frauke. Recht der Digitalisierung. Baden-Baden: Nomos, 2024.
MEDAUAR, Odete. O direito administrativo em evolução. 3. ed. Brasília: Gazeta Jurídica, 2016.
MORAES, Alexandre de. Democracia e redes sociais: o desafio de combater o populismo digital extremista. Rio de Janeiro: GEN Atlas, 2025.
MOREIRA NETO, Diogo de Figueiredo. Direito regulatório. Rio de Janeiro: Renovar, 2003.
SCHIRATO, Vitor Rhein. Direito administrativo brasileiro: genealogia, consolidação teórica, transformações e dogmática. São Paulo: JusPodivm, 2024.
SCHIRATO, Vitor Rhein. O manejo dos poderes fiscalizador e sancionador pela Administração Pública. Revista de Direito Administrativo Contemporâneo — ReDAC, São Paulo, v. 3, p. 41-71, nov./dez. 2013.
SCHIRATO, Vitor Rhein. O poder de polícia é discricionário?In: MEDAUAR, Odete; SCHIRATO, Vitor Rhein. Poder de polícia na atualidade. Belo Horizonte: Fórum, 2014.
SCHMIDT-ASSMANN, Eberhard. Das allgemeine Verwaltungsrecht als Ordnungsidee. 2. ed. Heidelberg: Springer, 2006.
SCHMIDT-ASSMANN, Eberhard. Das Verwaltungsrecht der Vereinigten Staaten von Amerika. Grundlagen und Grundzüge aus deutscher Sicht. In: Beiträge zum ausländischen öffentlichen Recht und Völkerrecht, Bd. 301. Nomos: Baden-Baden, 2021.
SCHOCH, Friedrich. Polizei-und Ordnungsrecht. In: SCHOCH, Friedrich (org.). Besonderes Verwaltungsrecht. 15. ed. Berlim: De Gruyter, 2013.
SILVA, Virgílio Afonso da. Direito constitucional brasileiro. São Paulo: Edusp, 2021.
SUNSTEIN, Cass R. Legal reasoning and political conflict. 2. ed. Oxford: Oxford University Press, 2018.
Article received and approved on February 2, 2026. DOI: https://doi.org/10.12660/rda.v285.2026.97522. ↑
Associate Professor of Administrative Law at the Law School of University of São Paulo (USP). Habilitated (Livre-Docente) and PhD in State Law from the Law School of University of São Paulo. Master of Laws (LL.M.) in Administrative Law from the University of Osnabrück. President of the Ibero-American Association of Regulation. Academic Secretary of the Center for Studies in Administrative, Environmental and Urban Law. Country: Brazil. E-mail: vrschirato@usp.br. ↑
Take, for example, the ongoing case brought by the US Federal Trade Commission against Meta Platforms for alleged monopolistic conduct and anticompetitive actions. Such litigation may, in extreme cases, lead to a judicial order for compulsory divestiture of assets, with the goal of fostering competition within the specific market. ↑
On this issue, see: MORAES, Alexandre de. Democracia e redes sociais: O desafio de combater o populismo digital extremista, Rio de Janeiro: GEN Atlas, 2025, p. 85 et seq. ↑
A clear example is the long-standing precedent set by the German Federal Constitutional Court, which designated the National Democratic Party (NPD) as fundamentally anti-democratic because of its overt ties to Nazi ideology. As a result, the party is banned and cannot participate in elections. See, in this regard, among other rulings: BVerfGE 104, 63-65. ↑
As Alexandre de Moraes observes, social media provides a fundamental link between extremist populist leaders and a disillusioned population, creating fertile ground for the spread of ideas that oppose democratic ideals and basic rights. On this point, see: Democracia e redes sociais: O desafio de combater o populismo digital extremista, p. 11 e ss. ↑
FUNK, William F.; SEAMON, Richard H. Administrative Law, 4th ed., New York: Wolters Kluver, 2012, p. 2. ↑
On this point, see: BOLDT, Hans; STOLLEIS, Michael. Geschichte der Polizei in Deutschland, pp. 9-12 ↑
See, for example: SCHOCH, Friedrich. Polizei-und Ordnungsrecht, in ______ (org.). Besonderes Verwaltungsrecht, 15ª ed., Berlim: De Gruyter, 2013, p. 168. ↑
On this point, see: SCHMIDT-ASSMANN, Eberhard. Das Verwaltungsrecht der Vereinigten Staaten von Amerika. Grundlagen und Grundzüge aus deutscher Sicht, in Beiträge zum ausländischen öffentlichen Recht und Völkerrecht, Bd. 301, Nomos: Baden-Baden, 2021, 420 S. ↑
On this topic, see, among others: SCHMIDT-ASSMANN, Eberhard. Das allgemeine Verwaltungsrecht als Ordnungsidee, 2. ed., Heidelberg: Springer, 2006, p. 139. ↑
On this topic, see my own work: O poder de polícia é discricionário? in MEDAUAR, Odete; _______. Poder de polícia na atualidade, Belo Horizonte: Fórum, 2021. p. 35 e ss. ↑
See, among others: ARAGÃO, Alexandre Santos de. O Poder Normativo das Agências Reguladoras, in Revista Trimestral de Direito Público n. 36, outubro/dezembro de 2001, São Paulo: Malheiros, p. 88 e seguintes. ↑
On the topic, see: DANWITZ, Thomas von. Was ist eigentlich Regulierung? Die Öffentliche Verwaltung, no. 23/2004, Stuttgart: Kohlhammer, 2004. ↑
Note that the concept developed here is a strict interpretation of regulation. It does not incorporate concrete measures the State may take, in the direct exercise of economic activities, to correct market failures. ↑
Beneath this spontaneous expression, however, lies a highly controlled system, wherein algorithms reinforce trends and direct consumers to suppliers. Key data is used to optimize connections between users and products that are more likely to appeal to them. Nevertheless, with the exception of a few outliers, there is little evidence of anticompetitive behavior. What we find is a tech-enabled sales strategy designed for maximum efficiency. ↑
On the topic, see: MARTINI, Mario; MÖSLEIN, Florian; ROSTALSKI, Frauke. Recht der Digitalisierung, Baden-Baden: Nomos, 2024, p. 49. ↑
In this context, many behaviors only recently addressed in the courts and codified as illegal have been socially taboo for much longer. It has long been recognized as offensive and undesirable to insult someone on the grounds of race (although it has always been, to some extent, relatively common). Yet, the criminalization of racial insults is fairly recent, only codified in Lei de Crimes de Preconceito (the Anti-Prejudice Law) in 2023, following the enactment of Federal Law No. 14,532 (in the wake of Federal Law No. 7,716/1989) ↑
For example, the German Federal Constitutional Court has already firmly ruled that making demonstrably false claims is not protected under the right to freedom of speech and expression. Thus, it provides a standard for adjudicating competing rights online, even if it carries some risks. On the issue, see: MARTINI, Mario; MÖSLEIN, Florian; ROSTALSKI, Frauke. Recht der Digitalisierung, p. 50-51. ↑
On the topic, see: SUNSTEIN, Cass R. Legal reasoning and political conflict, 2nd ed., Oxford: Oxford University Press, 2018, p. 192-193. ↑
Self-regulation is carried out by the social media support companies themselves, contractually, without the application of coercive state instruments. ↑
The approach was strongly advocated and, in theory, applied by tech companies after investigations stemming from the so-called Cambridge Analytica scandal, wherein it was found that a social network had allowed the use of data from more than 87 million users to create a system of influence over political decisions. ↑
Consider the recent case in which a Brazilian congressman posted falsehoods about the financial and tax systems, which was widely disseminated, attracting millions of views. In this case, the legality of deliberately lying to spread misinformation in this way is clearly an issue, as is the role of the social media platform in amplifying the comments. ↑
Consider, for example, the radical changes announced by one of the largest tech companies that controls three major social networks following the change of government in the US at the beginning of 2025. Under the new directive, filtering and removal of posts associated with hate speech is restricted, with an emphasis only on content associated with pedophilia and terrorism. ↑
Recently, social media regulation has been the subject of multiple cases before the U.S. Supreme Court. Two notable cases, Moody v. Netchoice and Netchoice v. Paxton, concerned the constitutionality of laws in the states of Texas and Florida, respectively, that prevented social media networks from removing certain content, and established rules for promoting and suppressing content. Ruling on these cases, the Supreme Court held that (i) social media users are protected by the First Amendment and thus their right to freedom of expression cannot be limited; (ii) the State cannot seek to limit the right to freedom of expression of citizens; and (iii) Social media platforms, per se, hold First Amendment rights and therefore have the power to moderate and remove content, and to decide what and how it may be amplified or suppressed. It is clear, then, from this brief summary that the prevailing understanding is that the right to freedom of expression cannot be limited, except in cases where there is proven harm to a third party. ↑
The principle of proportionality, established by the German Constitutional Court (Bundesverfassungsgericht) in the celebrated 1957 pharmacy trial. In the ruling of the court, any conflict between fundamental rights, which leads to the restriction of one, must be addressed on three fronts: necessity, adequacy, and proportionality, in the strict sense. On the topic, see: Among others, see: KINGGREEN, Thorsten; POSCHER, Ralf. Grundrechte — Staatsrecht II, 40th ed., Heidelberg: CF Müller, 2024, p. 101 et seq. ↑
Under the terms of Gesetz zur Verbesserung der Rechtsdurchsetzung in sozialen Netzwerken (Netzwerkdurchsetzungsgesetz— NetzDG) it is the responsibility of the Ministry of Justice to analyze the posted content and, if necessary, order its removal, in accordance with the procedures contained in the aforementioned law. ↑
On this topic, see, for example, the role played by the Brazilian Superior Electoral Court. Cf. MORAES, Alexandre de. Democracia e redes sociais—O desafio de combater o populismo digital extremista, p. 143 e ss. ↑
Clearly, the separation between police power and regulation, in the approach adopted here, is not watertight and limited. A state action may utilize both regulation and police power. ↑
The issue of strict legality in police powers in Brazil is a topic of continued debate. It seems clear that a sub-legislative normative act may establish regulatory conditions for the exercise of police power, provided that express legal authorization exists and clearly outlines the boundaries of that subordinate regulation. On the topic, see my own work: Direito Administrativo brasileiro — genealogia, consolidação teórica, transformações e dogmática. Thesis presented for the title of Livre-Docente in Administrative Law, USP, Faculty of Law, 2024, p. 597 et seq. (mimeo). ↑
On the means of inspection, see my work: O manejo dos poderes fiscalizador e sancionador pela Administração Pública, in Revista de Direito Administrativo Contemporâneo—ReDAC, vol. 3, novembro/dezembro de 2013, São Paulo: Revista dos Tribunais, 2018. p. 551. ↑
On the topic, see: MARTINI, Mario; MÖSLEIN, Florian; ROSTALSKI, Frauke. Recht der Digitalisierung, p. 52. ↑
Typical examples are the organs that operated during the Military Dictatorship (1964 to 1985), such as Conselho Nacional de Censura, in the Divisão de Censura de Diversões Públicas (DCDP) and Sistema Nacional de Informações (SNI). These had, by law, the authority to proscribe artistic, intellectual, and other expressions before their dissemination, which required prior approval by the relevant agencies concerned. ↑
On the topic, see: SILVA, Virgil Afonso da. Direito Constitucional brasileiro, São Paulo: EDUSP, 2021, pp. 171-172 and 174-175. ↑
But such an opinion should not be confused with the construction of a system to undermine the democratic rule of law: the systematic attack on democratic institutions and organization of means to overthrow them. ↑
This approach is used, for example, in the enforcement of smoking bans in enclosed spaces in the State of São Paulo, According to State Law No. 13541, of May 7, 2009, establishments are responsible for monitoring citizens, while the State oversees the businesses. ↑
Article 19 of the Brazilian Internet Bill of Rights provides that (in verbis): “Art. 400. To ensure freedom of expression and prevent censorship, internet application providers may only be held civilly liable for damages arising from third-party content if, following a specific judicial order if they fail to take steps to remove or disable access to the infringing material within the scope and technical limits of their service and the designated timeframe, unless otherwise provided by law. ↑
X (formerly twitter) is a clear example of this. When it was acquired by the current owner, there was widespread publicity that all control mechanisms would be eliminated, opening it up to any form of expression. As a result, many users of other networks whose content had been removed due to self-regulation policies migrated to X. ↑
Legally speaking, anyone who commits an unlawful act is obligated to compensate the victim. This, clearly, should encompass all those who contributed to the act, according to their level of culpability. Therefore, in the case of social media posts, the person who publishes the material and those who amplify it are liable to some extent. Article 19 of Marco Civil da Internet, therefore, contradicts the normative basis of civil liability in Brazilian law. ↑
The law provides that anyone who contributes to the commission of a crime must be held responsible for it, to the extent of their culpability. When the post constitutes a crime (for example, that of racism), both the author and the network which disseminates it, are to some degree criminally liable, Hence, the inherent contraction. ↑
Over recent decades, there has been a distinct increase in cases of strict civil liability within Brazilian law, expanding the range of circumstances from which the duty to indemnify arises. For example, Article 37, § 6 of the Federal Constitution establishes the objective liability of the State and its representatives. Similarly, Article 14, § 1 of Federal Law No. 6.938/1981 provides for liability in environmental law, while Article 12 of the Consumer Protection Code (Federal Law 8.078, 11 September 1990) mandates the liability of providers in relation to their consumers. ↑
The ruling on Theme 987 was as follows (in verbis): “The Court, by majority vote, considering topic 987 of general repercussion, denied the extraordinary appeal, with Justice Edson Fachin dissenting. Next, by majority vote, the following thesis was established: “Recognition of the partial and progressive unconstitutionality of article 19 of the MCI. 1. Article 19 of Law No. 12,965/2014 (Marco Civil da Internet), which requires a specific court order to hold an internet application provider civilly liable for damages arising from third-party content, is partially unconstitutional A state of partial legislative omission exists because the general rule established by Article 19 does not afford adequate protection to essential constitutional values, namely the preservation of fundamental rights and democratic principles Interpretation of Article 19 of the MCI 2. Until new legislation is enacted, Article 19 of the MCI (Brazilian Internet Law) should be interpreted in such a way that application providers are subject to civil liability, without prejudice to the specific provisions of electoral legislation and normative acts issued by the TSE (Superior Electoral Court). 3. The internet application provider will be held civilly liable, under the terms of article 21 of the MCI, for damages resulting from content generated by third parties in cases of crime or illegal acts, without prejudice to the duty to remove the content. The same rule applies in cases of accounts reported as inauthentic. 3.1. In cases of crimes against honor, Article 19 of the MCI applies, without prejudice to the possibility of removal by extrajudicial notification. 3.2. In cases of repeated replication of an offensive act already recognized by a court decision, all social media providers must remove posts with identical content, regardless of new court decisions, upon receiving judicial or extrajudicial notification. Presumption of responsibility 4. Providers are presumed liable for illegal content in relation to (a) paid advertisements and boosted posts; or (b) artificial distribution networks (chatbots or robots). In these cases, liability may be established regardless of notification. Providers will be exempt from liability if they can prove that they acted diligently and within a reasonable time to make the content unavailable. Duty of care in cases of widespread circulation of serious illegal content. 5. The internet application provider is liable when it fails to ensure the immediate removal of content that constitutes the commission of serious crimes as defined in the following exhaustive list: (a) anti-democratic conduct and acts that fall under the types provided in Articles 286 (sole paragraph), 359-L, 359-M, 359-N, 359-P, and 359-R of the Criminal Code; (b) crimes of terrorism or preparatory acts of terrorism, as defined by Law No. 13.260/2016; (c) crimes of inducement, instigation, or assistance to suicide or self-harm, pursuant to Article 122 of the Criminal Code; (d) incitement to discrimination based on race, color, ethnicity, religion, national origin, sexuality, or gender identity (homophobic and transphobic conduct), as categorized under the [respective] articles. 20, 20-A, 20-B and 20-C of Law No. 7.716/1989; (e) crimes committed against women by reason of their gender, including content promoting hatred against women (Law No. 11.340/06; Law No. 10.446/02; Law No. 14.192/21; Criminal Code, art. 141, § 3; art. 146-A; art. 147, § 1; art. 147-A; and art. 147-B of the Criminal Code); (f) sexual crimes against vulnerable persons, child pornography and serious crimes against children and adolescents, pursuant to articles 217-A, 218, 218-A, 218-B, 218-C, of the Penal Code and articles. 240, 241-A, 241-C, 241-D of the Statute of Children and Adolescents; g) trafficking in persons (CP, art. 149-A). 5.1 The liability of internet application providers provided for in this item concerns the configuration of a systemic failure. 5.2 Failure to adopt adequate measures to prevent or remove the illegal content listed is considered a systemic failure of the internet application provider, constituting a violation of the duty to act responsibly, transparently, and cautiously. 5.3. Measures are considered adequate if, according to current industry standards, they provide the highest levels of security for the type of activity performed by the provider. 5.4. The existence of illicit content in an isolated, atomized form is not, in itself, sufficient to give rise to application of civil liability under this item. However, in this case, the liability regime provided for in Article 21 of the MCI will apply. 5.5. In the cases outlined in this item, the party responsible for publishing the content removed by the internet application provider may legally request its reinstatement, provided they demonstrate the absence of illegality. Even if the content is restored by court order, the provider will not be required to pay compensation. Application of Article 19 6. Article 19 of the MCI applies to (a) email service providers; (b) providers of applications whose primary purpose is to hold closed video or voice meetings; (c) instant messaging service providers (also called private messaging service providers), exclusively with regard to interpersonal communications, protected by the confidentiality of communications (Article 5, item XII, of the Federal Constitution of 1988). Marketplaces 7. Internet application providers that operate as marketplaces are civilly liable under the Consumer Protection Code (Law No. 8.078/90). Additional duties 8. Internet application providers will be required to develop self-regulation that necessarily includes a notification system, due process, and annual transparency reports regarding extrajudicial notifications, advertisements, and boosted posts. 9. They should also provide users and non-users with specific customer service channels, preferably electronic, which are accessible and widely publicized on their respective platforms on a permanent basis. 10. These rules should be published and reviewed periodically, in a transparent and publicly accessible manner. 11. Internet application providers operating in Brazil must establish and maintain a headquarters and a representative in the country, whose identification and contact information must be readily available on their respective websites. This representation must grant the representative, necessarily a legal entity with its registered office in the country, full powers to (a) respond to administrative and judicial proceedings; (b) provide the competent authorities with information regarding the provider's operation, the rules and procedures used for content moderation and complaint management through internal systems; transparency reports, monitoring and management of systemic risks; rules for user profiling (where applicable), advertising and paid content boosting; (c) comply with judicial orders; and (d) respond to and comply with any penalties, fines and financial charges incurred by the represented party, especially for non-compliance with legal and judicial obligations. Nature of responsibility 12. There will be no strict liability in the application of the thesis stated here. Appeal to the legislator 13. The National Congress is urged to draft legislation capable of remedying the deficiencies of the current system regarding the protection of fundamental rights. Modulation of temporal effects 14.To preserve legal certainty, the effects of this decision are modulated, and it will only apply prospectively, except for decisions that have become final and unappealable. Ministers André Mendonça, Edson Fachin, and Nunes Marques were partially overruled on the argument. The opinion will be recorded by Minister Dias Toffoli (Rapporteur). Presidency of Minister Luís Roberto Barroso. Plenary session, June 26, 2025. ↑
See: KINGGREEN, Thorsten; POSCHER, Ralf. Grundrechte — Staatsrecht II, p. 86 et seq. ↑
While there is considerable controversy surrounding the effective origin of Administrative Law, its birth is properly traced back to Lei do 28 de Pluviose, Year VIII (February 17, 1800), as only then was a public executive power structured in subordination to the law. On this, see: MEDAUAR, Odete. O Direito Administrativo em evolução, 3ª ed., Brasília: Gazeta Jurídica, 2016, pp. 1-3. ↑
See MOREIRA NETO, Diogo de Figueiredo. Direito regulatório, Rio de Janeiro: Renovar, 2003, p. 107 et seq., especially 112 et seq. ↑
Some argue that racism or homophobia, for example, may only be defined as such if specifically directed at a particular individual or group of individuals. But, evidently, this is not the case. According to the Supreme Federal Court, any expression with racist content constitutes a crime and is therefore punishable. See, in this sense, HC 82424, reported by Justice Moreira Alves, (ruled on September 17, 2003, by Justice Maurício Corrêa) in which the Court held that an act of publishing antisemitic material was a crime of racism. ↑
For the sake of objectivity, we will not discuss the propriety of the STF establishing the extensive and detailed regulatory framework found in the Theme 937 ruling. This course of action is accepted based on the premise that a constitutional mandate provides the competence for such a measure. ↑