Brazilian Courts and the Internet – Rulings Before and After the Marco Civil on Intermediary Liability

Authors: Carlos Affonso Souza and Ronaldo Lemos
Institute for Technology & Society at Rio de Janeiro State University

Abstract: This paper discusses the treatment of online intermediary liability by Brazilian Courts both before and after the establishment of the Marco Civil. It first provides an overview of the three most common approaches to intermediary liability applied by the Courts in the decades prior to the Marco Civil. The paper then describes the drafting and passage of the Marco Civil – “The Brazilian Internet Bill of Rights” – as well as the system of civil liability for online intermediaries established by this document. While the Marco Civil has both clarified the position of online intermediaries in Brazil in regards to third-party content and established more robust protections for these entities, it is still too early to tell what the full implications of the implementation of the Marco Civil will be for the Brazilian Internet landscape.

Table of Contents

I. Introduction
II. Who is the Provider?
III. Brazilian Case Law on Liability of Online Intermediaries
A. The Provider Is Not Liable for the Conduct of Its Users
B. Strict Liability
C. Fault-Based Liability
1. General Effects of a Notice and Takedown Regime
D. The Special Case of "Search Providers"
IV. The Civil Liability of Online Intermediaries in the Marco Civil
A. Access Providers
B. Application Providers
1. Judicialization and Its Effects
2. Two Exceptions to the Liability Regime
i. Copyright
ii. Revenge Porn
V. Conclusion

I. Introduction

The Internet is a network that fosters freedoms, yet at the same time allows unprecedented control over individuals. It is an extraordinary platform for freedom of expression and – perhaps for the very same reason – can generate large-scale damage to one’s reputation and privacy.

This multitude of paradoxes guides the way to a deeper understanding of the dilemmas that must be addressed in order to reach a balance between the various interests of companies involved in the provision of access to the internet, and others services throughout this network.

Who is liable for damages caused online? The individual who posts a photo, a video, or a text that damages others, or the provider that, through its own activities, may provide the platform for that offense to occur?

This report focus on the current state of this discussion in Brazil, analyzing solutions created by a decade of judicial decisions on the topic of online intermediaries’ liability and the newly established regulatory framework as set forth in the so-called Marco Civil.

II. Who is the Provider?

Before exploring the debate around the framework for the civil liability regime for online intermediaries, it is necessary to identify exactly who an intermediary is in the Brazilian context.

Several authors have suggested different categories to identify intermediaries based on the activities they undertake. The Brazilian Superior Court of Justice (STJ), in assessing frequent cases involving damage over the Internet, has adopted the following classification:

"Internet service providers are those that offer several services related to the operation of the network. There are several categories of Internet service providers: (i) backbone providers, which hold the infrastructure capable of processing large volumes of information. They are responsible for Internet connectivity, offering its infrastructure to third parties, which end up contracting with end users to allow access to the network; (ii) access providers, which acquire access to the infrastructure through backbone providers and resell to end users, enabling them to have access to the Internet; (iii) hosting providers, responsible for the storage of third party’s data, allowing them remote access to it; (iv) information providers, which actually create the information available on the Internet; and (v) content providers, who make available the information created by information providers or by Internet users."[2]

Law No. 12965/2014 (known as "Marco Civil da Internet", or simply “Marco Civil”), deals in particular with two types of intermediaries: those dedicated to providing Internet access (“connection providers” or “access providers”) and those that provide the most diverse services to the network (“application providers”). Article 5 of this Law defines the activities performed by each of these providers.

Article 5. For the purposes of this Act, the following concepts apply: V - Internet connection: the enabling of a terminal for sending and receiving data packets over the Internet through the assigning or authentication of an IP address; (...) VII - Internet applications: the set of functionalities that can be accessed through a terminal connected to the Internet.

Understanding which activities the providers and their respective technical features perform is paramount to assessing the corresponding liability regime. In this regard, it is especially relevant to analyze the activities developed by their users and the extent to which the intermediary intervenes in these activities.

The question is thus how to create a liability regime that – on the one hand – does not harm the victim of any damage sustained online though the perpetuation of the illicit content, but which also does not encourage the intermediary to simply remove the photo, video, or text as soon as a notification is received, thus affecting the freedom of expression and diversity of online speech.

Is it up to the intermediary to judge the legality of the content displayed and to decide on its maintenance or removal? How do these questions impact the degree of innovation and future business models that may be affected by the way in which the law encourages prevention and imposes liabilities for damages caused online?

[2] STJ, Resp no. 1316921/RJ, Justice Nancy Andrighi; 26.06.12.

III. Brazilian Case Law on Liability of Online Intermediaries

After a decade of judicial decisions on the liability of online intermediaries in Brazil, one of three understandings were typically applied by national Courts. The first understanding exempts the provider from any liability for a third party’s behavior. The second enforces a strict liability regime for the Internet providers, grounded in the concept of the risk of the activity or in the recognition that a service was defectively rendered. A third and final understanding would link the liability of the provider to the existence of fault on its part. Some consider the provider liable simply for the non-removal of the content after the provider becomes aware of its existence (usually through a notification sent by the victim), while some understand liability as arising from non-compliance with a Court decision ordering the removal of the offending material. The latter was the understanding adopted in the recently enacted Marco Civil.

More than simply categorizing the understandings shared by national Courts, it is necessary to understand what the grounds are for supporting each position adopted by the Courts and which factual circumstances may have been relevant for the outcome of these decisions.

A. The Provider Is Not Liable for the Conduct of Its Users

The first understanding – according to which the provider would not be held liable for the acts of its users – is frequent in cases that identify the provider as a mere intermediary between the user (the offender) and the victim. In general there would be no conduct by the provider that would make it responsible for the acts of others. The provider’s only responsibility would be to help identify the offender.

In the majority of judicial decisions enforcing this understanding, it was very clear that the intermediary, in providing a specific service, was already informed that it would not be held responsible for content generated by its users, whether they are members of a social network or users of a webhosting platform.

In the beginning of the last decade, some Courts adopted this understanding, usually excluding the provider as a possible defendant in compensation claims filed by the victims of damages caused by users of the intermediary.

The Court of Appeals of the State of Paraná, analyzing a case involving offenses to the reputation of a victim made available on a website hosted by the intermediary, stated that:

"In the hosting agreement, the provider is only responsible for making an online space available. It must not interfere with the content that is published, except in cases of notorious illegality. The current Brazilian legal system does not allow the liability of the hosting provider, either strict or jointly, for moral damages arising out from the insertion of offensive material by the subscriber."[3]

Around the same period, a case involving a couple of lawsuits brought by former soccer player Paulo Roberto Falcão against Terra (an access and application provider) received some press coverage. The lawsuits claimed damages for offensive material that have been published on the website of the “Isto É Gente” magazine, hosted by Terra. The Court of Appeals of the State of Rio Grande do Sul recognized that the hosting provider could not be sued for the contents of a webpage that it simply hosts.[4]

[3] TJPR, Civil Appeal nr. 130075-8, 19.11.2002. [4] TJRS, Agravo de instrumento nr. 70003035078, Judge Paulo Antonio Kretzmann, 22.11.2001.

B. Strict Liability

The application of a strict liability regime for Internet providers is usually grounded in the notion of risk or defect in rendering a service to consumers. Regardless of the grounds adopted for its enforcement, the use of such an understanding in national case law leads to two relevant concerns.

First, does the provider have the duty to inspect, monitor, and consequently filter out content that is submitted by its users? That would be the very first concern, since the discussion around the duty of monitoring is key to understanding the effects of strict liability for providers. In this sense, the provider could be held liable for the mere display of harmful content (either because it is an inherent risk of their activity, or because there were a defect in the rendered service).

Second, should the provider be liable if, once aware of the reputedly harmful content – usually when notified by the victim – it does not remove it? This second concern takes into account the fact that providers should not be held liable for simply making the content available, but rather for the decision (active or passive) to not remove the challenged material.

Analyzing the first concern, it is relevant to highlight how the concept of risk has been applied in national case law. The large amount of lawsuits over damages caused online, especially focused on the use of social networks,[5] has drawn attention from the Judiciary due to the frequency in which such services are used to infringe third parties’ rights.

At the same time that access to the Internet in Brazil began to be widespread in the last decade, a new Civil Code was approved in 2002, providing for – in article 927 – strict liability for those who develop activities that, by their nature, involve risk to third party’s rights.

Therefore, as the number of cases brought to Court grew in the last decade, case law ended up establishing the understanding that a number of agents – from companies that operate search engines to owners of Internet cafes – could be held liable for the risk assumed in the development of their respective activities.

The Court of the State of São Paulo decided, in a lawsuit filed by the victim of defamatory messages sent our from an Internet café that the owner of such establishment should be liable as per "(...) the strict liability clause provided for in article 927, sole paragraph, of the Civil Code, as the development of its activity involves risk to the rights of others. (…) In this sense, whoever provides computer terminals or wireless network for Internet usage assumes the risk of misuse of the system to infringe third party’s rights, as it have happened in the present case."[6]

In opposition to such an understanding, a number of scholars sought to investigate not only whether there is risk in the activity, present in everyday situations, but also whether the risk posed by the activity performed by the intermediary is greater than usual. In this regard, the strict liability provision should only be applicable in extraordinary cases. As Erica Barbagalo explains:

"We understand that the activities undertaken by service providers on the Internet are not risk activities by their very nature, involving risks to rights greater than the one of any other commercial activity. Interpreting the law in the sense that any damage should be compensated regardless of culpability element would definitely burden the productive activities and therefore hinder development."[7]

The strict liability understanding based on the notion of risk was a minority view in the Court system by the end of the decade. The STJ, on several occasions, rejected this understanding. As stated in Special Appeal nr. 1308830/RS:

"The material damage resulting from messages with offensive content on the site uploaded by users does not constitute risk inherent to the activity of content providers, so that it does not apply to them the strict liability regime provided for in art. 927, sole paragraph, of the Civil Code."[8]

A second ground for the liability of the providers then lies in the characterization of the legal relationship between the victim and the intermediary as a true consumer relationship, which therefore results in application of the strict liability regime under the Consumer’s Protection Code (CDC).

After some debate in the late nineties on the enforcement of the CDC for online activities, it is worth noting that the main argument initially presented by providers when attempting to avoid liability was the non-essential nature of the service they were rendering.

Although a large numbers of services are rendered through the Internet without charging a specific value, national Courts decided that there is a counterpart offered by the consumer, even if it is of indirect nature. According to such case law, instead of paying a certain amount of money to the provider, this entity it earns profits from its users in other ways, especially through the creation of a user’s profile, which contains personal information and browsing habits, and can be used to generate advertising revenues through customized marketing based on the user’s data.

As explained by Professor Claudia Lima Marques:

"The expression used by art. 3 of the Consumer Protection Code include all consumer services rendered in connection to a ‘remuneration’ of some sort. (...) It seems to me that the choice of the expression 'paid' creates an important opportunity to include consumer services paid indirectly, ie, when it is not the individual consumer who pays, but the collectivity or when he/she pays indirectly. The term 'compensation' allows you to add all those contracts in which it is possible to identify the hidden synallagma (hidden counterpart), an indirect remuneration for the service rendered to the consumer." [9]

The subject was addressed a number of times in Superior Court of Justice decisions. In one of the first cases to reach the STJ on the liability of Internet service providers, the Court stated that, "to characterize the consumer relationship, the service can be provided by the provider for remuneration obtained indirectly." [10]

More recently, the STJ reinforced this understanding, thus confirming the enforcement of the CDC in the relationship between the provider of a social network and the victim of offenses made available in a community created on Orkut:

"Commercial use of the Internet is subjected to the regulation of consumer relations arising out from Law No. 8.078/90. 2. The fact that the service provided by the Internet service provider is to be free does not change the nature of the relationship as a consumer one, since the term remuneration, contained in art. 3, §2, ​​of the CDC, should be interpreted broadly so as to include the indirect gain from the provider."[11]

Once the relationship is subsumed under the Consumer Protection Code, it remains to ascertain whether the damage caused by the service can be framed as a defect in the service. The question of risk appears again to reveal the importance of the first concern mentioned above: if the provider has a duty to monitor the content that is made available on its pages, the mere display of harmful content implies a defect in the service rendered.

The STJ has already decided on several occasions that the service provider has no obligation to monitor the content of text, photos, videos, and codes entered by its users. As stated in the Special Appeal nr. 1308830/RS:

"A prior inspection on the content of the information posted on the web for each user by the content provider is not intrinsic to the service, so it can not be deemed a defective activity under article 14 of the CDC, if the site does not examine and filter the data and images uploaded by its users."[12]

Another opposing argument to the imposition of surveillance duties (and consequently to the strict liability understanding) is the assertion that, by requiring the inspection of the posted contents, a censorship regime would be implemented, hindering freedom of expression.

This argument is represented by the Republic’s General Attorney in a currently ongoing case to be decided by the Supreme Court (STF) involving the creation of a community in the Orkut social network that was reputedly offensive. The community “I hate Aliandra” (“Eu odeio a Aliandra”) was created to mock a high school teacher in the State of Minas Gerais. The teacher then filed a suit against Google for the damages caused by this content.

The lawsuit questions if the providers should monitor what is said in the community pages created on social networking sites as a way to prevent future damage. According to General Attorney:

"…There is no interference from the provider in the content posted by users on the social networks, being incompatible with the constitutional framework to allow or even to require previous censorship of disseminated manifestations, under penalty of strict liability. It would amount to undue and severe embarrassment to the very freedom of expression."[13]

The STJ, in support of this view, has even claimed that, "prior editorial control of the content of the information equates to a breach of confidentiality of correspondence and communications, prohibited by Art. 5, XII, of the Constitution."[14]

The understanding of the STJ of service providers in general needs to be analyzed carefully because, on one hand, the Court recognizes that such relationships are subject to the Consumer Protection Code but, on the other, it does not impose a strict liability (as it would be the rule in the CDC). If there is a negative answer to the first concern raised (i.e. that "providers have no duty to monitor and are not liable simply by making a content available"), then there is a need to examine the second, which require an investigation regarding whether the provider would be liable if it fails to remove the infringing content once it has become aware of its existence.[15]

[5] In May 2012, Justice Nancy Andrighi mentioned that around 200 lawsuits involving Google alone were pending decision in the STJ (STJ, Resp 1308830/RS, Justice Nancy Andrighi; 08.05.2012).
[6] TJSP, Process nr. 583.00.2006.243439-5, Judge Ulysses de Oliveira Gonçalves Junior; 06.03.2008.
[7] Erica B. Barbagalo. “Aspectos da Responsabilidade Civil”, in Ronaldo Lemos, Ivo Waisberg (orgs) Conflito de Nomes de Domínio e Outras Questões Jurídicas da Internet. São Paulo: RT, 2003; p. 361. See also STJ, Resp 1067738/GO, Justice Sidnei Beneti, 26.05.2009.
[8] STJ, Resp 1308830, Justice Nancy Andrighi; 08.05.2012. See also STJ, RESP 1306066/MT, Justice Sidnei Beneti; 17.04.2012.
[9] In Comentários ao Código de Defesa do Consumidor. São Paulo: Revista dos Tribunais, 2003; p. 94.
[10] STJ, Resp 566468/RJ, Justice Jorge Scartezzini, 23.11.2004.
[11] STJ, Resp 1308830/RS, Justice Nancy Andrighi; 08.05.2012.
[12] STJ, Resp 1308830/RS, Justice Nancy Andrighi; 08.05.2012. See also STJ, Resp 1316921/RJ, Justice Nancy Andrighi; 26.06.2012.
[13] Manifestação da Procuradoria Geral da República, RE nr. 660861/MG, 11.07.2012.
[14] STJ, Resp 1308830/RS, Justice Nancy Andrighi; 08.05.2012.
[15] See STJ, Resp 997993/MG, Justice Luis Felipe Salomão; 21.06.2012.

C. Fault-Based Liability

The third understanding of intermediary liability is based on the existence of a fault by the provider, attaching to itself the responsibility for the conduct performed by its user. This understanding has two different grounds for application: the first states that the liability should result from noncompliance with a notification informing the provider of the infringing material; the second is based on noncompliance with a Court order requesting the removal of certain material. This last understanding is the one adopted by the Marco Civil.

In its most recent decisions on the issue, the STJ affirmed the understanding that Internet service providers can be liable when they fail to remove illegal content of which they are aware by a notification sent by the victim. Such an understanding has been enforced for both cases in which the provider fails to respond to the notification of the victim, or actively responds to the notification stating that it sees no reason to remove the content. In such cases, the responsibility would be based on fault and jointly affirmed with the liability of the user that has directly uploaded the infringing material.

On this topic it is worth mentioning some relevant excerpts from Special Appeal nr. 1.193.764/SP, as decided by the STJ:

"By offering a service through which it allows users to freely express their opinions, the content provider should take care to provide resources so that it can identify those users, curbing anonymity and assigning each event a certain authorship. From the perspective of the average diligence expected of the provider, it must adopt the measures which, according to the specific circumstances of each case, can individualize the website’s users, under penalty of being liable for fault (culpa in omittend)."[16]

This debate was also reported in the decision of the STJ in the Regimental Appeal presented in the Special Appeal nr. 1.309.891/MG. In this case there is a deeper discussion over the expression "immediate" as to how quickly the provider should act to remove the infringing content:

"In line with precedents of this Court, the Internet content provider is not liable according to a strict liability regime for the content created by the user in a website, as this is not an inherent risk to their activity. It is required, however, to immediately takedown the morally offensive content, otherwise it would be jointly responsible with the direct offender. Precedents. In the present case the Court held that there was no immediate exclusion of the fake profile because the victim, for more than once denounced the illegality perpetrated by electronic means provided for this purpose by the provider itself, without obtaining any result. Regimental Appeal dismissed."[17]

The case above, as decided by the STJ, verified the decision of lower level Court that found Google – when exploring the social network Orkut – was not diligent in promoting the removal of offensive material as it took eleven days to remove the content. Cases like this call into question the frequent use of the expressions "immediately" or even "energetic" by the STJ when it comes to damages caused through online intermediaries.

[16] STJ, Resp 1193764/SP, Justice Nany Andrighi; 14.12.2010.
[17] STJ, Agr. Reg. in Resp 1309891/MG, Justice Sidnei Beneti; 26.06.2012.

1. General Effects of a Notice and Takedown Regime

The liability of the provider for not removing the content once notified seems intuitive: if the provider is aware that someone claims to be suffering damages due to a content made available by your user, the one who stands in the best condition to cease the damages – other than the offender himself – would be the provider. However, this hides many harmful consequences for the operations of the Internet and for the protection of many fundamental rights.

At first one must question whether the provider should promptly remove the content and thus prevent the ongoing damage. Would it be appropriate for the provider to analyze whether the content is or isn’t actually infringing? The danger of this alternative lies in empowering providers to decide what should and what should not be made available on criteria that go beyond those presented in their terms of service.

The STJ has had the opportunity to express some concerns with this broad delegation of the power to control speech online to private actors. As mentioned by Justice Nancy Andrighi:

"We must consider the impossibility to define a criteria that would authorize the veto or the disposal of given page. Given the subjectivity surrounding the psychological damage and/or the damage to one’s image, it would be impossible to define parameters that could allow the providers to rely on to define whether a content is potentially offensive. On the other hand, it would be reckless to delegate this judgment to the discretion of the providers."[18]

The second point worth mentioning is precisely the intense subjectivity of the criteria that can be used to allow content to be removed. If it does not make sense to hold the providers liable just because content was made available and there is doubt on whether it is infringing or not, then a system that lacks transparency and that is highly subjective, removing content and jeopardizing the diversity and the degree of innovation on the Internet, should undoubtedly be rejected.

The degree of innovation on the Internet is the third point that can be mentioned in opposition to a system of fault-based liability arising out from the failure to remove content after being notified. The development of all new activity involves questioning its adherence to the current legal regime and, in most cases, an investigation into potential judicial decisions on the subject. The removal of content in a very subjective way and by a mere notification creates serious obstacles to the development of new alternatives for exploration and communication on the Internet, dampened by fear of future claims that could be filed if notifications requiring the removal of contents are not "immediately" complied with.

A fourth important point relates to judicial analysis of cases that could provide greater legal certainty for business developed on the Internet. If, for fear of liability, providers end up taking down massive amounts of content, the immediate result is a reduction in the number of cases on which the Judiciary could act to draw the limits of expression in the Internet. This could relegate the establishment of mechanism for content removal to private parties, resulting in processes that might not be in accordance with judicial standards for expression in other media, for instance.

A notice and takedown regime that renders the provider liable for not taking down certain content after being notified creates two alternatives equally detrimental to the diversity of the discourse on the network. Either the provider takes down the content as soon as it receives the notification and thus gives rise to the whole range of abuses stemming from the ease of removing content that may be harmful to others (with strong impact on freedom of speech, of press), or the provider fights to maintain the content online, understanding that it has no reason to be removed and thus assuming the risk of being held legally responsible for that very content. This situation creates little incentive to protect freedom of expression for all providers and creates a strong disincentive for small providers that cannot bear the burden of mass litigation.

Therefore, even if the application of liability based on fault offers superior results to that obtained by imposing strict liability, one must realize that to affirm liability arising from the failure to comply with a notification has a number of negative implications for the way in which the Internet operates. Thus, this system needs to give way to liability rooted in compliance with Court decisions, such as that provided by the Marco Civil.

[18] STJ, Resp 1316921/RJ, Justice Nancy Andrighi; 26.06.2012.

D. The Special Case of "Search Providers"

Before dealing with the liability regime provided by the Marco Civil itself, it is worth noting that the Superior Court of Justice (STJ) has given different treatment with regards to liability to so-called "search providers" than to other services and applications, such as social networking and webhosting. According to recent decisions by the STJ, when acting as a simple search engine Google will not be held liable for the content displayed as search results.

The most famous case that affirms this understanding is a case involving the actress and TV host Xuxa Meneghel, which sought to compel Google to remove from the engine all results for the search term “pedophile xuxa” or even other result involving the name of the Plaintiff, partially or fully written, regardless of spelling, in connection to any criminal act."[19]

The motivation for the lawsuit was the widespread availability on the Internet of an early 80s movie called “Love Strange Love” (“Amor Estranho Amor”),[20] in which the actress is featured in two scenes having intimate relations with a 12-year old boy. Much of the actress’ concern is due to the fact that, shortly after the movie was released, she began a career on TV hosting a show focused on kids and teenagers. The availability of such material online could hamper the image she has created in recent decades.

The STJ decided in favor of Google in this lawsuit. The decision is grounded in the relevance of search providers (part of the “application providers” category, in the language of Law nr. 12695/14) in indexing the information found on the Internet. According to the leading vote of Justice Nancy Andrighi:

“Search providers perform their searches within a virtual universe, whose access is public and unrestricted, ie, its role is limited to the identification of web pages where certain data or information, even if illegal, are being freely made available. Thus, although the search engine facilitate access and the consequent dissemination of pages whose content is potentially illegal, these pages are public as parts of the world wide web and therefore appear as results of the research sites.”[21]

As a consequence of the role played by search providers on the Internet, they can not be required to overturn the indexing mechanism for addressing third parties’ pages; this would unduly interfere with the legitimate collective interests, such as access to information. According to the judgment:

“Search providers should not be required to remove from their system the results derived from the search term or expression, nor the results that point to a specific photo or text, regardless of the indication of the URL of the page where it is inserted. It is not advisable, for the purposes of hindering the spread of illegal or offensive content on the web, to suppress the right to information. Balancing the rights involved and the potential risk of violation of each of them, the odds should favor the guarantee of freedom of information, as set forth in Art. 220, § 1, of the Constitution, especially considering that the Internet is today an important vehicle of mass media.”[22]

Therefore, the decision from STJ indicates that the victim should seek to prosecute whoever is responsible for the damage, such as the person who actually published the illegal content, and refrain from prosecuting the search provider that only indexes the information freely found on the web.

“If the conditions for the exclusion of a certain webpage are found, under the allegation of unlawful or offensive content - notably the identification of the URL of this page - the victim will lack interest to act against the search provider. If the victim identified though the URL the author of the illegal act, it has no reason to sue the one who merely facilitates access to this act, which has been so far publicly available on the network.”[23]

Two comments seem especially relevant on cases involving search providers in the STJ. The first concerns to the difference in treatment accorded to the search engines compared to the liability regime typically adopted for social networks and video hosting sites. There are Court decisions that not only require providers to indemnify the content, but which also require these intermediaries to remove content in accordance with specific instructions from the Plaintiffs, creating a permanent channel for the exclusion of content based on the requirements authorized by the Court when a request is made.

The STJ, in tackling the case of the search provider, explicitly rejected this possibility, as mentioned in the previous judgment by the Court of Appeals of the State of Rio de Janeiro (TJRJ). As the vote of Justice Nancy Andrighi details:

"Finally, it is important to assess the feasibility of the solution adopted by TJRJ, creating a process for removal of a certain content, previously indicated by the victim. This form of restriction, if applicable, should always arise out from a judicial order, as it would be impossible for a simple extrajudicial notification to achieve such result. To have this process conducted through private notifications would end up delegating the judgment about the offensive potential of a given text or image to the discretion of the victim or the provider. At the same time, there are precedents from this Court involving similar cases - liability for the content of offensive messages on social networks – in which we have decided, in general terms, that "once notified that a certain text or image is unlawful, the provider must act energetically to remove the material immediately. Failing to do so would render the provider jointly liable with the direct offender" (Resp 1.186.616/MG, DJE 31.08.2011. In the same direction, see: REsp 1.193.764/SP, DJE of 08.08.2011). In the specific case of social networks, the intermediary itself provides a system for complaints on its own platform, allowing users to report unlawful or offensive content. The respective term of use gives the provider the right to remove any page or content that is in breach of the Term of Service. Therefore there is a special agreement, which authorizes the provider to exercise a discretionary judgment, a circumstance that is absent in search engines. The use of such applications does not even require registration. It is essential, therefore, that the request for exclusion from the search results of a particular text or image is made ​​in Court."[24]

The second comment relates to the targeting by the victim of a specific search provider, instead of others that could be used to find the exact same content. In this case, the market share of Google results in a situation in which the company finds itself as the defendant in the overwhelming majority of the lawsuits against search providers.

This dilemma is no stranger to the STJ. Recent decisions highlighted the paradox and limitations of civil liability regimes when applied to the Internet. According to the vote of Justice Nancy Andrighi in Special Appeal No 1407271/SP:

"…It must be noticed that [the victim] acted exclusively against Google when the video can be found through the use of several other search engines. Consulting CADÊ and BING sites, for example, held by MICROSOFT and YAHOO companies respectively, we have been able to find more than 100,000 results for the same term."[25]

Would the victim then be obliged to file a suit against all, or at least against the most relevant search providers to show the seriousness of his/her complaints? If this seems a strange requirement to protect one's rights to reputation, image, and privacy, it shows how lawsuits against search providers may not serve the best interest of the victim. It is increasingly difficult to achieve the total removal of harmful material, especially in the current stage of technological progress, with the constant emergence of new ways to share pictures, videos, and text from mobile devices and the large availability of ways to play, download, store, and encrypt content.

[19] STJ, Resp 1316921/RJ, Justice Nancy Andrighi; 26.06.12.
[20] http://en.wikipedia.org/wiki/Love_Strange_Love.
[21] STJ, Resp 1316921/RJ, Justice Nancy Andrighi; 26.06.2012.
[22] STJ, Resp 1316921/RJ, Justice Nancy Andrighi; 26.06.2012.
[23] STJ, Resp 1316921/RJ, Justice Nancy Andrighi; 26.06.2012.
[24] STJ, Resp 1316921/RJ, Justice Nancy Andrighi; 26.06.2012.
[25] STJ, Resp 1407271/SP, Justice Nancy Andrighi; 21.11.2013.

IV. The Civil Liability of Online Intermediaries in the Marco Civil

Law nr. 12965/2014 seeks to establish "principles, guarantees, rights and obligations for the use of the Internet in Brazil" according to its first article. This Law is the result of a pioneer initiative, led by the Brazilian government, to use the Internet as a pool for consultation on the actual content of forthcoming legislation. Even before arriving in the National Congress, during the online debate phase of the initiative the issue of intermediary liability was one of the most debated topics, along with net neutrality and data protection.

The so-called Marco Civil (or “Brazilian Internet Bill of Rights”) was the first experience in Brazil with the use the Internet as a way to broaden the discussion of a Draft Bill of Law, ensuring that a much more significant number of participants could get involved in the debate of the legislation.

Specifically concerning the civil liability regime for intermediaries, Law nr. 12965/14 provides two different treatments depending whether the intermediary falls into the category of connection/access provider or application provider.

A. Access Providers

Holding the access provider liable for the acts of its users is a practice that has been rejected by national and international Courts since the late nineties.[26] There are two common arguments for recognizing the non-liability of connection providers for the damages caused by third parties that are simply using their services to connect to the Internet.

The first argument lies in the technical impossibility on the part of providers to avoid harmful behavior by its users. It is noteworthy that this preventive conduct by connection providers is not only impossible but also undesirable, since it would lead inevitably to an increase in mass surveillance practices of controversial legality.

The second argument transcends the technological aspect by focusing on the rupture of any nexus (“nexo causal”) between the damage caused to a third party and the act of simply providing network access to a user. The simple Internet connection does not seem to be the direct and immediate cause of the damage suffered by a victim, rather the damage is caused by the behavior specifically played out by the user that generated the illegal content.

Law nr. 12965/14 echoes such arguments in Article 18, as it exempts connection providers from liability for the actions of its users:

Article 18. The provider of connection to Internet shall not be held liable for civil damages resulting from content generated by third parties.

It is important to mention that the exemption set forth in Article 18 only applies to cases in which the provider would be held liable for third party conduct. Connection providers are still liable for the damages they cause directly through their own activities, as shown by a large pool of cases decided in the national Courts. Among the cases involving the liability of connection providers are situations involving damage to their own users, such as the failure to provide services dully contracted by their users or in the different conditions than the ones previously established by either contract or the relevant sectoral regulations.

[26] See Religious Technology Center v. Netcom On-Line Communication Services, Inc, 21.11.1995. In Brazil, among many decisons, see: TJRS, Ap. Civ. n° 70001582444, Judge Antônio Correa Palmeiro da Fontoura, 29.05.2002.

B. Application Providers

The liability of application providers is provided in the Article 19 of the Marco Civil in the following terms:

Art. 19. In order to ensure freedom of expression and to prevent censorship, the provider of Internet applications can only be subject to civil liability for damages resulting from content generated by third parties if, after an specific Court order, it does not take any steps to, within the framework of their service and within the time stated in the order, make unavailable the content that was identified as being unlawful, unless otherwise provided by law.

§ 1. The referred Court order must include, under penalty of being null, clear identification of the specific content identified as infringing, allowing the unquestionable location of the material.

§ 2. The implementation of the provisions of this article for infringement of copyright or related rights is subject to a specific legal provision, which must respect freedom of speech and other guarantees provided for in art. 5o of the Federal Constitution.

§ 3. The compensation disputes for damages arising from content made available on the Internet related to the honor, reputation or personality rights, as well as the removal of related contents by Internet application providers, can be presented to special small causes Courts.

§ 4. The judge, including within the proceeding set forth in § 3º, can anticipate, partially or in full, the effects of the request contained in the initial petition, to the extent that undisputable proof exists of the fact, considering society’s collective interest in the availability of the content on the Internet, as long as the requisites of truthiness of the author’s claims, the reasonable concern of irreparable damage, or damage that is difficult to repair are met.

As previously mentioned, the Marco Civil affirms that the general rule for intermediary liability in Brazil is based on the fault of the provider, denying the attempts to hold them liable in typical strict liability standards, either by the simple availability of harmful content based on the risk theory or based on the rendering of a defective service. At the same time that the Marco Civil evades strict liability, the approach it provides for liability based on fault is quite different from the usual liability arising out from the simple lack of action after being notified that damages are being caused by the availability a certain material.

Here lies perhaps one of the most heated controversies of the Law: the Marco Civil provides that intermediaries are only held liable if they fail to fulfill a Court order requesting the removal of content.

One of the most frequent criticisms of this provision is that the Marco Civil only allows content to be removed by a Court order. This is not the best interpretation of the mentioned provision. What the Marco Civil sets forth is a safeguard for application providers in the sense that they will only be held liable if they do not comply with a Court order requesting the removal of the offensive material. This provision does not prevent intermediaries from determining their own requirements for removing content once notified by the alleged victims of damages arising out from materials made available through their platforms.

The Marco Civil gives freedom of expression high importance in this debate, guaranteeing to the providers an immunity that neutralizes any concern that they would have of being held liable for a lack of content removal once notified.

As mentioned by André Zonaro Giacchetta, analyzing the text while on debate in the National Congress:

"The text of the Draft Bill clearly favors the guarantee of the rights of Internet users, instead of restricting their liberties. This is a standard created for the user in good faith. There is a clear choice for ensuring freedom of thought and expression, as well as the privacy of Internet users and the protection of personal data."[27]

Additionally, it is noteworthy that the solution provided by Law nr. 12965/14 does not necessarily oblige the victim to file a lawsuit in order to have the content removed. Such removal will depend on the terms of service of the website, the nature of the infringing content, the persuasive language of the notification in evidencing the damages caused by the same material and etc. However, the Marco Civil directs the settlement of any dispute between the victim and the provider to the Judiciary, as it recognizes Judiciary Power as precisely the legitimate authority to solve the controversy.

[27] André Zonaro Giacchetta. “A Responsabilidade Civil dos Provedores de Serviços de Internet e o Anteprojeto de Reforma da Lei n 9610/98 (“Lei de Direitos Autorais”)”, In Revista da Associação Brasileira da Propriedade Intelectual, n. 117 (mar-abr/2012); p. 39.

1. Judicialization and Its Effects

The Marco Civil fosters the understanding that an intermediary should not be compelled to remove content simply because a notification has been received. The provision of Article 19, as mentioned above, creates incentives for the claim to be brought to the Judiciary.[28]

One recurrent argument in this regard is the fact that the speed in which contents might be copied and shared through the Internet is not compatible with the time it takes for a lawsuit to be brought to the Judiciary. At the same time, it is important to stress that the Marco Civil expressly provides that a judge may order the removal by granting the victim an injunction in cases when it seems clear that the delay in taking the content down would worsen the victim’s situation.[29]

In order to make this solution easier and faster for the victim of certain damages, the Marco Civil states that such cases can be brought to the Special Small Claims Courts. The provision in the third paragraph of Article 19 makes reference to cases of “compensation disputes for damages arising from content made available on the Internet related to the honor, reputation or personality rights, as well as the removal of related contents by Internet application providers.”

The balance that the Marco Civil tries to achieve aims at accommodating all interests involved, protecting freedom of expression by clearly defining the role of the provider and ensuring that they must play a prominent role in the prevention and elimination of damage, while avoiding arbitrary judgments or fear of future liability.

If the situation is brought to a Court, the Marco Civil recognizes the Judiciary as the most appropriate forum for the resolution of such cases. At the same time, an interesting side effect of the Marco Civil is that it fosters capacity building of judges on the evolution of modern technologies for information and communication, as such knowledge is crucial to the exercise of their functions.

In affirming that application providers must only be held liable in cases in which fault is found, and not by simply failing to comply with a notification, the Marco Civil separates itself from the case law that has been created in the last decade in Brazil, especially by the Superior Court of Justice.

[28] See Marcelo Thompson. “The Insensitive Internet – Brazil and the Judicialization of Pain“ (http://www.iposgoode.ca/wp-content/uploads/2010/05/Marcelo-Thompson-The-Insensitive-Internet-Final.pdf).
[29] See Marcel Leonardi. Responsabilidade Civil dos Provedores de Serviços na Internet. Brasília: Juarez de Oliveira; p.207.

2. Two Exceptions to the Liability Regime

Law nr. 12965/14 has two important exceptions to the general liability regime, as described in the article 19: copyright infringement, as provided by the second paragraph of the article, and cases of so-called "revenge porn," provided by Article 21.

For both cases the general rule that intermediaries may only be held liable if they fail to comply with a Court order demanding the removal of the content is not applicable. These two situations, for very different reasons, can trigger the liability of the provider if it is notified and fails to remove a specific content.

i. Copyright

The exception concerning copyright was due to a continuous demand, especially by radio and television broadcasters, for the Marco Civil not to change the established practice of sending out notifications for the removal of copyrighted material made available without proper authorization or in circumstances not protected by the exceptions and limitations regime as set forth by the Copyright Act (Law 9.610/98). Brazilian Courts have recognized several times the liability of the application provider when, once notified, it fails to remove the content.

An additional circumstance that explains why such an exception was inserted in the review process of the original text of the Marco Civil in the National Congress was the fact that the Federal Government, through the Ministry of Culture, has been developing in recent years a process of consultations for the reform of the Copyright Act, dealing with topics such as liability for copyright infringements carried out online.

In this regard, the removal of further considerations on liability through copyright infringement would prevent the existence of two different regimes for the very same issue in Brazil: the one in the Marco Civil and the other as provided by an eventual reform of the Copyright Act.

It is worth noting that the Marco Civil has not simply deferred the treatment of such matters to the Copyright Act. The second paragraph of Article 19 of Law No. 12965/14 states that the regulation of online copyright infringement should be tackled by the Copyright Act, but at the same time it states that treatment under this Act should "respect freedom of speech and other guarantees provided for in Article 5 of the Federal Constitution."

The final part of this provision is quite revealing since one of the guidelines of the reform of the Copyright Act is to achieve a better balance between Copyright and other fundamental rights, such as access to knowledge and freedom of expression, and at the same time preventive abusive conduct in copyright enforcement. In this sense the Marco Civil advances some of the concerns of the Copyright Act reform, as envisioned by the Ministry of Culture, already providing an interpretive clause for whichever solution is adopted in the reform of the specific law.

ii. Revenge Porn

The second exception to the rule in Article 19 of the Marco Civil is the provision of Article 21 for cases of so-called "revenge porn"[30] materials.

The provision was inserted in one of the last rounds of editing on the text of the Bill and it was clearly motivated by the suicide of two Brazilian girls after intimate adult videos end up being shared through Whatsapp. A number of Congressmen have referred to this case as the trigger for creating an exception to the general rule on intermediaries’ liability.

Art. 21. The Internet application provider that makes third party generated content available shall be held liable for the breach of privacy arising from the disclosure of images, videos and other materials containing nudity or sexual activities of a private nature, without the authorization of the participants, when, after receipt of notice by the participant or his/hers legal representative, refrains from removing, in a diligent manner, within its own technical limitations, such content.

Sole Paragraph. The notice set forth above must contain sufficient elements that allow the specific identification of the material said to violate the right to privacy of the participant-user and the confirmation of the legitimacy of the party presenting the request.

Article 21 creates a different liability regime from the general rule of Article 19 for the cases in which the application provider fails to remove materials that fall into the category presented above. It is important to highlight that the final part of the provision makes this exceptional liability conditional on evidence that the provider(s) have not acted in a diligent manner. This condition – together with the addition of the expression “within its own technical limitations” – could provide an opportunity for discussion in the forthcoming lawsuits on what the standards should be for how providers should act when they are given notice that intimate material, such as the ones targeted by this provision, has been made available through their applications.[31]

[30] http://pt.wikipedia.org/wiki/Pornografia_de_vingan%C3%A7a.
[31] See STJ, Resp nº 1306157/SP, Justice Luis Felipe Salomão, 17.12.2013.

V. Conclusion

After more than a decade of case law on the liability of online intermediaries, the enactment of Law nr. 12965/14 tries to balance all relevant interests in the development of several online activities. The so-called “Marco Civil da Internet” has, since its origin, been intended to establish a human rights oriented perspective for the regulation of the Internet in Brazil.

Whether such balance has been achieved is a debate that still depends on how Brazilian Courts will interpret its provisions, enforce fundamental rights such as privacy, data protection, and freedom of expression, all in connection with the need to respect such rights and – at the same time – create a fruitful environment for innovation and development online.

Such an initiative is not entirely comprehended without knowledge of international and domestic politics around the negotiation of some of its most relevant provisions. In this case, the fallout from the Snowden revelations and a presidential election campaign had a great impact on how the text of the Marco Civil came to be.

In regards to the intermediary liability provisions, the addition of two exceptions – for copyright and revenge porn – during the last year of negotiations in the National Congress offers a glimpse of how different stakeholders have organized themselves for the protection of their respective interests in this piece of regulation.

As Brazil bridges the digital divide, the Marco Civil will serve as umbrella legislation, setting the principles for future regulation on matters concerning the Internet. As more and more people, especially from the peripheries of Brazil, connect to the network, it will be interesting to follow up on how practices and behaviors change.

The influx of Brazilian users in platforms intended for global usage, such as Google’s Orkut, has not only resulted in very innovative uses by Brazilians, but also a significant opportunity for balancing different interests in Internet regulation. The immense pool of judicial decisions on damages caused by Orkut’s users is a complex and not entirely explored body of research material.

Hopefully this report has covered the most relevant judicial decisions concerning online intermediary liability in Brazil and can serve as a guide to navigate the intriguing future of a country which, after a decade of debate, has finally enacted a human rights oriented piece of legislation that aims to promote the values that are inherent in the current stage of Internet development, while at the same time providing room for diversity and innovation.