Image Rights and False Claims, Data Protection and Intermediary Immunity: the case of Russmedia

 

 

Lorna Woods, Professor Emerita, University of Essex

 

Image credit: US Department of Defense

 

This Grand Chamber judgment of the Court of Justice in X v Russmedia Digital and Inform Media Press (Case C-492/23) handed down on 2 December 2025 concerns the scope of data protection rights and intermediary immunity in the context of the non-consensual use of someone’s image.  The judgment identifies:

- when someone has responsibilities under the GDPR,

- the relationship between those regulatory obligations and intermediary immunity, and

- the steps an data controller could take to satisfy those GDPR obligations.

 

It has been described as reshaping the obligations of online operators in the EU, while others have questioned how far the points in the judgments may be generalised to other situations.

 

Judgment

 

The Facts

 

Russmedia owns an online marketplace on which advertisements may be published. An unidentified user posted an advertisement falsely representing X as offering sexual services. The advert included X’s photographs (though there is no suggestion that these were intimate images) and phone number, all without her consent. Once notified, Russmedia removed the advert within an hour but the advertisement had been shared across several third party websites and remained accessible. X sued in the national courts in respect of her image rights, rights to reputation and data protection rights. The Romanian courts struggled with the question of whether Russmedia could claim the benefit of intermediary immunity (under the e-Commerce Directive (Directive 2000/31), provisions now replaced by the Digital Services Act (DSA)) and the extent of the obligations under the GDPR.

 

Is Russmedia subject to Obligations under the GDPR?

 

Obligations under GDPR arise when (1) personal data are (2) processed by (3) a data controller.

 

The CJEU commenced its analysis by noting the the information contained in the advert about X was personal data for the purposes of the GDPR and moreover that claims about a person’s sex life (implied in the advert) constituted “sensitive” personal data as protected by Article 9 GDPR, and that remained the case whether or not the claim was true.  Classification of the data as special category data means that there is a higher threshold to show lawful processing of those data. 

 

The Court further noted that “the operation of loading personal data on a webpage constitutes processing” for the purposes of the GDPR (para 54) and therefore covered the publication of the advert.

 

While this puts the advert within the scope of the GDPR, its obligations apply to data controllers and processors, so the question was whether, given Russmedia had no control over the content of the advert, was it a controller or joint controller? The Court reiterated previous jurisprudence to say (para 58) that:

 

any natural or legal person who exerts influence over the processing of such data, for his or her own purposes, and who participates, as a result, in the determination of the purposes and means of that processing, may be regarded as a controller in respect of such processing.

 

It noted also that there may be more than one entity which is a controller in respect of processing – this is the idea of joint controllers, although they may not have equal responsibility depending on the facts (para 63).  Joint decision making is not necessary for there to be joint controllers.

 

While the test for “controller” requires that the person processing the data does so for their own purposes, the Court added that this could include the situation “where the operator of an online marketplace publishes the personal data concerned for commercial or advertising purposes which go beyond the mere provision of a service which he or she provides to the user advertiser”  (para 66).  The Court in this case pointed to the fact that the terms of use give Russmedia “considerable freedom to exploit the information  published on that marketplace” including “the right to use published content, distribute it, transmit it, reproduce it, modify it, translate it, transfer it to partners and remove it at any time” (para 67). Russmedia is therefore not publishing solely on behalf of the user placing the advert. The Court also noted that Russmedia make the data in the advert accessible, allows the placing of anonymous adverts and sets the parameters for the dissemination of adverts (likely to contain personal data).

 

As a result of finding that the advert publishing platform was a joint controller the GDPR obligations bite in relation to the advert and must be able to demonstrate that the advert is published lawfully, which includes the requirement for consent for sensitive data (para 84 and 93) and the requirement for accuracy.  The CJEU notes that once published online and accessible to any Internet user, such data may be copied and reproduced on other websites, so that it may be difficult, if not impossible, for the data subject to obtain their effective deletion from the Internet.  The adds to the seriousness of the risks facing the data subject.

 

The GDPR also requires the implementation of technical and organisational measures – and this should be considered in the design of the service so that such data controllers can identify adverts containing sensitive data before they are published and to verify that such sensitive data is published in compliance with the principles of the GDPR (para 106).  Further, the controller must ensure that there are safety measures in place so that adverts containing sensitive data and not copied and unlawfully published elsewhere (para 122).

 

Are the GDPR Obligations Affected by Intermediary Immunity?

 

While the immunity provisions in the e-Commerce Directive are far-reaching, the e-Commerce Directive specified that it was not to apply to the Data Protection Directive (the legislation in  force at the time the e-Commerce Directive was drafted) and that included the immunities; the Court concluded that this meant the e-Commerce Directive could not interfere with the GDPR. It also specified that GDPR requirements here cannot be classified as general monitoring (which is prohibited by the e-Commerce Directive (and now the DSA)).

 

 

Conclusions, Implications and Questions

 

The ruling in this case does not match existing industry practice. It is not a bolt out of the blue, however, but builds on existing jurisprudence (eg Fashion ID (Case C-40/17)).  While the obligations required of Russmedia in this case may indicate, to some, a landmark shift in the Court’s approach, the judgment does rely on the specific facts in the case and, specifically, the point that “sensitive” data, which effectively requires explicit consent, is in issue. In principle, this could be relevant to other forms of sensitive content, notably non-consensual intimate images (NCII). Certainly, it re-emphasises data protection as a route for victims’ redress, if not preventing harm in the first place.

 

The ruling clarifies that a range of activities typically carried out by platforms - structuring, categorizing, and monetizing user content, can amount to determining “the purposes and means of processing personal data”, the test for responsibility as a controller under the GDPR (article 4 GDPR). In taking this approach, it differed from the Opinion of its Advocate-General (AG’s Opinion, para 120).  The Court noted that the definition of controller in the GDPR is broad – and this is to support the protection of individuals’ fundamental rights to privacy and data protection. Once a body is a controller, that body must be able to demonstrate compliance with the data protection principles, and take appropriate technical and organisational measures to ensure data processing is carried out in accordance with the GDPR. 

 

Here, some of the points that the Court relied on to determine that Russmedia was a joint controller could well be relevant to other services and not just online marketplaces. For example, many sites have broad terms of service similar to those the Court highlighted here; other services also allow anonymous posting and a key feature of many services is the making available of that content for advertising revenue purposes, as well as controlling how content is promoted. (Note the decision of the court in YouTube and Cyanado (Joined Cases C-682/18 and C-683/18), which suggested that automated content curation did not mean that a service is not neutral, is not directly relevant here as it relates to the conditions for maintaining intermediary immunity – and see Russmedia, AG’s Opinion, para 155)  It is unclear how many of these criteria need to be present for a service to constitute a controller in relation to the personal data in third party content it publishes (though the Court seems to list them as alternatives, suggesting any of them would suffice), or whether less far-reaching terms of service may be sufficient to stop a platform being a joint controller.  Where these conditions are satisfied, its impact need not be limited to advertising but to organic content containing third party personal data too.

 

The Court’s confirmation that the clear wording of the e-Commerce Directive, excluding the Data Protection Directive (the predecessor legislation to the GDPR) from its scope, meant that an intermediary cannot escape its own data protection responsibilities does not affect immunity from liability in respect of unlawful content.  Note that this decision was based on the wording of the e-Commerce Directive. This language has not been carried over to the DSA, which is expressed to operate without prejudice to, inter alia, the GDPR. It is not clear if or how this would change the Court’s interpretation. Immunity provisions from the e-Commerce Directive have been carried across to the DSA (albeit with a “carve out” in respect of consumer law in Article 6(3) DSA). While the EDPB has published guidance on the interplay of the GDPR and the DSA, it has looked at the question of the impact of the DSA requirements on data protection rather than the impact of data protection on the DSA.

 

The judgment suggests that services should design checks into their services to ensure compliance with the data protection obligations including pre-publication checks as to whether sensitive data is included and to check the identity of the person posting the material. Of course, while some sorts of posts (eg NCII) clearly constitute sensitive personal data, the outer edges of this category might not be clear cut. The Court here noted that the category should be interpreted broadly (para 52). It could be that some of the obligations could be passed on to the user uploading the advert through terms of service, though this might be capable of being abused by some users.  Further, the CJEU expects the site to prevent third party scraping so far as is possible – the judgment does not introduce strict liability in this regard.  What technical measures would be sufficient in practice remains uncertain.   This is very different from the reactive response required to maintain immunity under the e-Commerce Directive – and which has been the dominant framing until now. Assuming the position on immunity does not change, services may have to implement new systems, probably including automated tools and may ultimately affect choice of business model for some services.

 

There are questions about how this ruling impacts the DSA. How does a pre-check system differ from general monitoring. General monitoring is prohibited under Article 8 DSA (though specific monitoring is not)? The CJEU stated that systems to ensure GDPR compliance could not be classified as “general monitoring” (para 132) – but did not explain this statement any further. There is an argument to say that all content will need to be scanned to identify that which contains sensitive personal data – and by contrast to checking against a database of known CSAM images, for example, which might be considered specific monitoring, this is a more open ended obligation. It is unclear whether there are other routes to pre-check which do not involve content scanning.  The requirements to check whether the person posting the personal data is the person to which the data relates (or is otherwise lawfully processing) may make, for example, anonymity difficult to maintain and it is unclear what level of identity verification would be acceptable.  There are also questions about how this system of pre-checks affects the neutrality of the platform and consequently the possibility for the platform to claim immunity (in respect of other claims relating to the content) under Article 6 DSA.

 

The position in the UK may be slightly different, however. Section 6(1) European Union (Withdrawal) Act provides that decisions of the CJEU post-dating 31 December 2020 do not bind UK courts although they may have regard to such judgments. The provisions which would have had the effect of removing the status of binding precedent from decisions of the CJEU made on or before that date have now not been brought into force (but they remain on the statute book), as the Labour Government revoked the relevant commencement regulations.  Furthermore, old case law from the Northern Irish courts (pre-dating Brexit), CG v. Facebook, suggested the the e-Commerce Directive (the relevant law at the time) could apply to data protection claims.

Facebook’s liability for defamatory posts: the CJEU interprets the e-commerce Directive

Lorna Woods, Professor of Internet Law, University of Essex

The last couple of weeks have seen a number of judgments relating to the control of information on the internet by the subject of the information.  The cases of GC et al (Case C-136/17) and Google v CNIL (Case C-507/17) concern the interpretation of General Data Protection Regulation (GDPR), looking at the obligations of search engines. The most recent case, Glawischnig-Piesczek v Facebook (Case C-18/18) concerned the impact of the e-Commerce Directive (Directive 2000/31/EC), specifically the prohibition on general monitoring found in Article 15 of that Directive, on ‘stay down’ notices. The focus of this post is on Glawischnig-Piesczek, but there is a question that reaches beyond the impact of that case on the e-Commerce Direcitve: to what extent is there a coherent approach to issues arising from the Internet across the various legal measures that intersect with it. This may go beyond the e-Commerce Directive and the GDPR to include measures related to intellectual property (notably the recent controversial Directive on Copyright in the Digital Single Market (Directive 2019/790/EU) and the Enforcement Directive (Directive 2004/48/EC)) and the combating of child exploitation (Directive on combatting the sexual abuse and sexual exploitation of children and child pornography (Directive 2011/93/EU)) and terrorism (Terrorism Directive (Directive 2017/541/EU)).

The Judgment

The facts giving rise to this reference are simple.  Glawischnig-Piesczek complained to Facebook about some defamatory posts. Facebook did not remove the posts so Glawischnig-Piesczek obtained a court order requiring Facebook to stop publishing the impugned content. The precise scope of the order gave rise to further litigation and the Austrian Supreme Court referred a number of questions to the CJEU, asking:

-          Are “stay down” notices in relation to identically worded content compatible with Article 15 of the e-commerce Directive?

-          Are there geographic limitations to the obligation?

-          Are such notices in relation to content with equivalent content to that which has been found unacceptable which does not use the same words but conveys the same meaning acceptable?

-          In relation to posts with equivalent meaning, does the obligation accrue when the intermediary becomes aware of the content?

The Court started its analysis by making clear that the immunity from suit granted by Article 14 of the Directive is not a general immunity from every legal obligation. Specifically the national authorities remain competent to require a host to terminate access to or remove illegal information. The Court also noted that Article 18 of the e-Commerce Directive requires Member States to have in place appropriate court actions to deal with illegal content. It states:

Member States shall ensure that court actions available under national law concerning information society services’ activities allow for the rapid adoption of measures, including interim measures, designed to terminate any alleged infringement and to prevent any further impairment of the interests involved.

The Court held that no limitation on the scope of such national measures can be inferred from the text of the e-Commerce Directive [para 30]. 

It then turned to the impact of Article 15. It highlighted the fact that while Article 15 prohibited general monitoring as recital 47 in the preamble of the Directive makes clear, monitoring ‘in a specific case’ does not fall within that prohibition. It then held that

[s]uch a specific case may, in particular, be found, as in the main proceedings, in a particular piece of information stored by the host provider concerned at the request of a certain user of its social network ….. [35].

Given the nature of information flows there is a risk that any such information may be re-posted, so

‘.. in order to ensure that the host provider at issue prevents any further impairment of the interests involved, it is legitimate for the court having jurisdiction to be able to require that host provider to block access to the information stored, the content of which is identical to the content previously declared to be illegal, or to remove that information, irrespective of who requested the storage of that information. In particular, in view of the identical content of the information concerned, the injunction granted for that purpose cannot be regarded as imposing on the host provider an obligation to monitor generally the information which it stores, or a general obligation actively to seek facts or circumstances indicating illegal activity, as provided for in Article 15(1) of Directive 2000/31 [37].

The Court determined “equivalent meaning” to be about the message the information posted conveys and which was “essentially unchanged”. Given the focus on meaning not form, the Court held that an injunction could extend to non-identical posts as otherwise the effects of an injunction could easily be circumvented.  The Court then considered the balance between the competing interests. The Court commented that the “equivalent information” identified by court order should contain specific elements to identify the offending content and in particular must not require the host to carry out its own independent assessment. In terms of assessing the burden on the host, the court noted that the host would have recourse to “automated search tools and technologies” [para 46].

The court concluded that the injunctions would not constitute a general obligation to monitor all content and specifically no obligation to seek facts or circumstances indicating illegal activity.

The Court also noted that Article 18 of the Directive makes no provision for territorial limitations on what measures Member States may make available. In principle, world-wide effects would be permissible [para 50], but this is subject to the proviso that EU rules must be consistent with the international law framework.

The court felt it unnecessary to respond to the third question without elaborating further.

Comment

There are a number of comments that could be made about this judgment. This post comments on three: the approach of the Court to general monitoring; non-identical content; and the issue of territorial scope. It also discusses freedom of expression issues.

General Monitoring

In this judgment there is a clear confirmation that searching for specific pieces of information/types of content does not constitute general monitoring. The Court makes it clear, at para 35, that the searching for individual pieces of content constitutes a ‘specific case’ within recital 47.  The Court gives the searching for specific information as an example of a ‘specific case’; presumably the searching of a targeted user’s stored date could be another such.  This is the first time the approach has been adopted in regards to defamation.  Perhaps the fact that the case concerns defamation rather than, for example, intellectual property explains the dearth of previous case law cited in the court’s judgment.

The statement of the Court that searching for an individual item of content does not constitute general monitoring does not address the fact that such a search would presumably involve search all content held. Yet in McFadden (Case C-484/14), the Court described the scope of Article 15 thus:

As regards, first, monitoring all of the information transmitted, such a measure must be excluded from the outset as contrary to Article 15(1) of Directive 2000/31, which excludes the imposition of a general obligation on, inter alia, communication network access providers to monitor the information that they transmit. [McFadden, 87]

This is broadly similar to the approach in the early case of L’Oreal (Case C-324/09) which referred to Article 15 precluding ‘an active monitoring of all the data of each of its customers in order to prevent any future infringement ...’ [para 139].  The prevention of future infringements in the context of L’Oreal could be achieved – in the view of the Court – however by the suspension of the perpetrator.  Yet monitoring of the data of all customers seems to be what would be required to find the specific case of content. The matter remains unacknowledged in the Court’s analysis in Glawischnig-Piesczek.  Perhaps the assumption is (a) that the concern underpinning the prohibition in Article 15 derives from privacy; and (b) that when we look for one thing we do not really see the other things that we look at during our search – and that this might particularly be so when the searching is automated.  Of course, arguments that automated review of communications data does not constituted an invasion of privacy have not been accepted by the Court (e.g. Watson/Tele2).  In any event, further support for the distinction between searching for an identified piece of content and searching in a less targeted fashion is found in the context of the fight against child sexual abuse and exploitation and in the context of enforcement of intellectual property.

Non-identical content

The clarification that an injunction may also extend to non-identical content raises a number of issues.  While the Court states that a host should not be required to make its own judgment on these matters it is not clear how similar the content needs to be.  It is also unclear whether the Court is concerned here with the wording or the message conveyed. At [40] it refers to the ‘message conveyed’, which could refer to the idea in issue rather than its precise expression. The Court then referred to ‘information, the content of which, whilst essentially conveying the same message, is worded slightly differently …’ [41]. An approach based on wording (or presumably identifiable items of content such as images) has the benefit of being more easily described by order. It is probably easier to circumvent.  There seems to be an assumption in the judgment that technological means are available to implement this sort of requirement, though whether that is the case is another question.

Territorial Scope

The territorial scope of the order is also worth mentioning.  Like its Advocate General, Szpunar, the Court does not envisage any territorial limitation of the removal/blocking as a result of EU law.  It is important to note that the Court does not say that injunctions must have such extraterritorial effect. Rather the question is about the interplay between each national legal system’s own way of dealing with these issues (and area the Court noted gave Member States wide discretion) and the fact that Article 18 is silent as to any limitations. The silence of EU law allows Member States freedom to take action.  A further issue is that the Court noted the impact of international law without however elaborating what it meant – are we talking fundamental human rights (this seems unlikely given the existence of the EU Charter) or international comity, for example?. 

This is one of a number of cases in which the Court has had to consider the territorial scope of EU law in the context of the Internet – the most recent being the Right to be Forgotten case: Google v CNIL (Case C-507/17). There the Court held that

Where a search engine operator grants a request for de-referencing …, that operator is not required to carry out that de-referencing on all version of its search engine, but on the version of that search engine corresponding to all the Member States. [73]

It seems then that the opposite conclusion has been reached. This is overstating the point.  While the fact that EU law does not require extraterritoriality, the GDPR’s silence on the point gives space to a national court to make an order with extra-territorial effect, a point the Court makes express in para 72. A national could then, within the constraints of its own national rules, make such an order. In such a situation the position would seem similar to that under Article 18 e-Commerce Directive as understood in Glawischnig-Piesczek. A contrast to the silence of the EU legislature on extraterritoriality of blocking/de-listing can be seen in the Terrorism Directive. There Article 21 imposes an obligation on Member States to obtain the removal of terrorist content hosted outside their territory, but it also recognises that that may not be possible.

In Google v CNIL, while the Court recognised the possibility for national courts to make orders for de-referencing with extra-territorial effect, it expressly noted that in doing so they must weigh up the competing interests of the data subjects  and the right of others to freedom of information [para 72]. It is noticeable that in Glawischnig-Piesczek the balancing is different. The Court notes the interest of the subject of the information and also the need not to impose an excessive burden on the host provider [para 45, para 46]. The existence of other rights: the right of the host to carry on a business and the rights of those posting the material and those wishing to receive it – both aspects of freedom of expression - are not expressly mentioned. To some extent the issue of rights will be covered through the national courts, which will be the bodies to carry out that balancing within their own national frameworks and within the limits of EU law. By contrast to Google v CNIL, however, there is no instruction from the Court that these are matters to be considered, nor any express recognition that the balance between the right to private life (including the protection of reputation) and freedom of expression differs between territories. What might be seen as the legitimate protection of private life in one place is an infringement of speech in another.  So, while the matter was not directly the Court’s concern in this case, it is somewhat surprising that the issues were not directly considered.

Barnard & Peers: chapter 9

Photo credit: Department of Defense, via Wikicommons

Facebook, defamation and free speech: a pending CJEU case

Preliminary Notes on the Pending Case Glawischnig-Piesczek v. Facebook Ireland Limited

Dr Paolo Cavaliere, University of Edinburgh Law School, paolo.cavaliere@ed.ac.uk

Introduction

In the next few months the Court of Justice of the European Union is expected to deliver a decision with the potential to become a landmark in the fields of political speech and intermediary liability (the Advocate-General’s opinion is due June 4). In fact the Court will have to render its opinion on two intertwining yet distinct questions: first, the case opens a new front in the delineation of platforms’ responsibility for removing illegal content, focusing on whether such obligations should extend to content identical, or even similar to other posts already declared unlawful. Secondly, the decisions will also determine whether such obligations could be imposed even beyond the territorial jurisdiction of the seised court. What is at stake is, in ultimate analysis, how much responsibility platforms should be given in making proactive assessments of the illegality of third-party content, and how much power courts should be given in imposing their own standards of acceptable speech across national boundaries.

Summary of the case

The plaintiff is a former Austrian MP and spokeswoman of the Green Party who – before retiring from politics – was reported as criticizing the Austrian Government’s stance on the refugee crisis in an article published by the news magazine oe24.at in April 2016. A Facebook user shared the article on her private profile along with her own comment, which included some derogatory language. In July, the plaintiff contacted Facebook and requested the post to be removed, only for the platform to decline the request as it did not find the post in breach of its own terms of use nor of domestic law. The plaintiff then filed an action for interim injunctive relief seeking removal of the original post, and of any other post on the platform with ‘analogous’ content. After the Commercial Court of Vienna found the post unlawful, Facebook proceeded to remove it.

However, the Court considered that Facebook, by failing to remove the original post on the plaintiff’s first notice, was not covered by the exemption from secondary liability and ordered the platform to remove any further post that would include the plaintiff’s picture alongside identical or ‘analogous’ comments. The Higher Regional Court of Vienna then found that such an obligation would amount to an obligation of general monitoring on Facebook’s part and removed the second part of the injunction, while upholding that the original post was manifestly unlawful and should have been removed by the platform following the first notification from the plaintiff. The Higher Court also confirmed that Facebook should remove any future posts that would include the same derogatory text alongside any image of the plaintiff. Facebook appealed this decision to the Austrian Supreme Court.

The Court referred to the CJEU two main ranges of questions:

- First, whether it would be compatible with Article 15(1) of the E-commerce Directive an obligation for host providers to remove posts that are ‘identically worded’ to other illegal content. In case of positive answer, the Court asks whether this obligation could expand beyond identical content and include content that is analogous in substance, despite a different wording. These are ultimately questions concerning the responsibility that platforms can be given in making their own assessment of what content amounts to unlawful speech, and what are the limits of “active monitoring”.

- Second, whether national courts can order platforms to remove content only within the national boundaries, or beyond (‘worldwide’). This is a question concerning the admissibility of extra-territorial injunctions for content removal.

Analogous content and active monitoring

To start with, it needs to be clarified that the dispute focuses effectively on a case of political speech, only formally concerns a case of defamation. The post on Facebook was considered by the Austrian court in breach of Art 1330 of the Austrian Civil Code, which protects individual reputation. However, the status of the plaintiff, who served as the spokeswoman of a national political party at the time, gives a different connotation to the issue. Established case-law of the European Court of Human Rights (Lingens v. Austria, 1986; Oberschlick v. Austria (no. 2), 1997) has repeatedly found that the definition of defamation in relation to politicians must be narrower than usual and the limits of acceptable criticism wider, especially when public statements susceptible of criticism are involved. In this case, the plaintiff had made public statements concerning her party’s immigration policy: the circumstance is relevant since the ECtHR traditionally identifies political speech with matters of public interest and requires interferences to be kept to a minimum. By established European standards the impugned content here amounts to political commentary, and the outcome of the case could inevitably set a new standard for the treatment of political speech online.

While intermediaries enjoy a series of immunities under the E-commerce Directive, which also notably established a prohibition for state authorities to impose general monitoring obligations, the 2011 Report of the Special Rapporteur to the General Assembly on the right to freedom of opinion and expression exercised through the Internet clarified that blocking and filtering measures are justified in principle when they specifically target categories of speech prohibited under international law, and the determination of the content to be blocked must be undertaken by a competent judicial authority. A judicial order determining the exact words (‘lousy traitor’, ‘corrupt oaf’, ‘fascist party’) may be an adequately precise guidance for platforms to operate, depending on how precise the contours of the order are.

To put the question in a context, the requirement to cancel ‘identical’ content marks the latest development in a growing trend to push platforms to take active decisions in content filtering. It cannot be neglected that the issue of unlawful content re-appearing, in identical or substantially equivalent forms, is in fact becoming increasingly worrisome. In a workshop held in 2017, delegates from the EU Commission heard from industry stakeholders that the problem of repeat infringers has become endemic to the point that, for those platforms that implement notice-and-takedown mechanisms, 95% of notices reported the same content to the same sites, at least in the context of intellectual property infringements. If rates of re-posting of content infringing other personality rights such as reputation can be considered anecdotally similar, then any attempts to clear platforms of unlawful content recall the proverbial endeavor of emptying the ocean with a spoon.

Nonetheless, the risk of overstepping the limits of desirable action is always looming. A paradigmatic example comes from early drafts of Germany’s Network Enforcement Law, which included a requirement for platforms to prevent re-uploads of same content already found unlawful – a provision that closely resembles the one at stake here. The requirement was expunged from the final version of the statute amid fears of over-blocking and concerns that automated filters would not be able, at the current state of technology, to correctly understand the nuances and context of content that is similar of equivalent at face value, such as irony or critical reference for instance.

The decision of the German law-makers to eventually drop the requirement – evidently considered a step too far even in the context of a statute already widely considered to hardly strike a suitable balance between platform responsibilities and freedom of expression – is indicative of the high stakes in the decision that the CJEU faces. A positive answer from the CJEU would mean a resurgence of this aborted provision on Europe-wide scale.

The idea of platforms’ monitoring of re-uploaded content is being gaining traction in digital industries for a little while now and is trickling down into content regulation. In the field of SEO, the concept of “duplicate content” defines content that has been copied or reused from other Web pages, sometimes for legitimate purposes (e.g. providing a mobile-friendly copy of a webpage), sometimes resulting in flagrant plagiarism. Yet definitions diverge when it comes to the criteria considered: while duplicate content is most commonly defined as ‘identical or virtually identical to content found elsewhere on the web’, Google stretches the boundaries to encompass content that is ‘appreciably similar’. Content regulation simply cannot afford the same degree of flexibility in defining ‘identically worded’ content, as the criterion of judicial determination required by the Special Rapporteur and the prohibition of general monitoring obligations in the E-commerce Directive exclude it.

In the area of copyright protection, it is in principle possible for service providers like YouTube to automatically scan content uploaded by private users and compare it to a database of protected works provided by rights-holders. In the case of speech infringing personality rights and other content-based limitations, discourse analysis is necessary to understand the context, and this kind of task would evidently amount to a private intermediary making a new determination on the legality of the speech.

The assessment of what amounts to unlawful speech rarely depends on the sole wording; context plays a fundamental role in the assessment, and that is all but a straightforward exercise. The European Court of Human Rights’ case-law includes several examples of complex evaluations of the local circumstances to determine whether or not an interference with speech would be justified.

For instance, in the case of Le Pen v. France (2010), the Court considered that comments, that could seem at face value derogatory towards a minority, needed anyway to be considered in the context of an ongoing general debate within the Country, and stressed that the domestic courts should be responsible for assessing the breadth and terms of the national debate and how to take it into account when determining the necessity of the interference. In Ibragimov v. Russia (2018), the Court noted that the notion of attack on religious convictions can change significantly from place to place as no single standard exists at the European level and, similar to political debates in societies, domestic authorities are again best placed to ascertain the boundaries of criticism of religions ‘[b]y reason of their direct and continuous contact with the vital forces of their countries’. The historical context is consistently taken into account to determine whether a pressing social needs exist for a restriction, and is enough to justify different decisions in respect to speech that otherwise appears strikingly similar.

For instance, outlawing Holocaust denial can be a legitimate interference in countries where historical legacies justify proactive measures taken in an effort to atone for their responsibility in mass atrocities (see Witzsch v. Germany (no. 1), 1999; Schimanek v. Austria, 2000; Garaudy v. France, 2003); whereas a similar statute prohibiting the denial of the Armenian genocide would be an excessive measure in a country like Switzerland with no strong links with the events in 1915’s Ottoman Empire (Perinçek v. Switzerland, 2015).

The intricacies of analysing the use of language against a specific historical and societal context are perhaps best illustrated by the Court’s minute analysis in Dink v. Turkey (2010). The Court was confronted with expressions that could very closely resemble hate speech: language such as ‘the purified blood that will replace the blood poisoned by the “Turk” can be found in the noble vein linking Armenians to Armenia’, and references to the Armenian origins of Atatürk’s adoptive daughter, were included in articles written by the late Turkish journalist of Armenian origin Fırat Dink.   The Court eventually came to the conclusion that it was not Turkish blood that Dink referred to as poison, but rather attitudes of the Armenian diaspora’s campaign which he intended to criticise. The Court built extensively on the assessment made by the Principal State Counsel at the Turkish Court of Cassation – who analysed all Dink’s articles published between 2003 and 2004 – in order to be able to ascertain whether these expressions amounted to denigrating Turkishness, and in what ways references to blood and the origins of Atatürk’s daughter amounted to sensitive subjects in Turkish ultranationalist circles and were susceptible to ignite animosity.

Not only social and political context matters, often it is precisely the use of language in a culturally specific way that forms a fundamental part of the Court’s assessment, with the conclusion that words alone have little importance and it is instead their use in specific contexts that determines whether or not they cross the boundaries of lawful speech. In Leroy v. France (2008), the Court went to great lengths in evaluating the use of the first person plural “We” and a parodistic quote of an advertising slogan to establish that a cartoon mocking the 9/11 attacks amounted to hate speech.  

Beyond the Court’s experience, examples of words that, though otherwise innocuous, can become slurs if used in a certain context abound: for instance, the term ‘shiptari’ in Southern Slavic-speaking countries to indicate Albanians especially in Serbia acquires a particularly nasty connotation as it was often used by Slobodan Milošević to show contempt of the Albanian minority in Yugoslavia. In Greece, the term lathrometanastes (literally ‘illegal immigrants’) has been appropriated and weaponised by the alt-right rhetoric to purposefully misrepresent the legal status of asylum seekers and refugees in an attempt to deny them access to protection and other entitlements, and now arguably lies outside of the scope of legitimate political debate,[1] to the point that it has been included in specialised research on indicators of intolerant discourse in European countries.

This handful of examples shows how language needs to be understood in the context of historical events and social dynamics, and can often convey a sense beyond their apparent meaning. While for domestic and supranational courts this seems challenging enough already, the suggestion that it would suffice for platforms to just check for synonyms and turns of phrases in a mechanical fashion is simplistic at best.

Extraterritorial injunctions

This plain observation calls into question whether it would be appropriate for the CJEU to answer in the positive the question on extra-territorial injunction. The Austrian Court’s order is in fact addressed towards an entity based outside the Court’s territorial jurisdiction and the order sought is to operate beyond the Austrian territory. To clarify, the novelty here resides not on the seising of Austrian courts, but rather on the expansive effect of their decision; the question concerns whether it would be appropriate for the effects of the injunction sought to extend beyond the limits of the national jurisdiction and effectively remove content from Facebook at the global level.

The Court of Justice has already interpreted jurisdiction in a similarly expansive way on a few occasions. In L’Oréal v. eBay (2011) the Court decided to apply EU trade-mark law on the basis that trade-mark protected goods were offered for sale from locations outside the EU but targeted at consumers within the EU. In Google Spain (2014), the Court decided to apply EU data protection law to the processing of a European Union citizen’s personal data carried out ‘in the context of the activities’ of an EU establishment even if the processor was based in a third country. The Court considered that delimiting the geographical scope of de-listings would prove unsatisfactory to protect the rights of the data subjects. A similar reasoning was the basis for deciding in Schrems (2015) that EU data protection law should apply to the transfer of personal data to the US.

One common element emerges from the case-law of the Court of Justice so far, in that the extraterritorial reach of court orders seems to be a necessary measure to ensure the effectiveness of EU rules and the protection of citizens’ or businesses’ rights. The Court has been prepared to grant extraterritorial reach when fundamental rights of European Union citizens were at stake (for instance in the context of processing of personal data) or when, in case of territorially protected rights such as trade marks, a conduct happening abroad was directly challenging the protected right within the domestic jurisdiction. It is dubious that the case at stake resembles either of these circumstances; in fact limiting political speech requires a different analysis.

A politician certainly is entitled to protect their own reputation, however when the criticism encompasses aspect of an ongoing public debate, the limits of acceptable speech broaden considerably: whether the speech falls within, and contributes to, an ongoing social conversation is very much a factual and localised consideration. Conversations that are irrelevant or even offensive within one national public sphere could very well be of the utmost relevance to communities based elsewhere, especially minorities or diasporas, who could find themselves deprived of their fundamental right to access information.

The CJEU has traditionally paid attention to connecting factors justifying extraterritorial orders. Following its own jurisprudence, it will now be faced with the challenge of identifying a possible connecting element to justify a worldwide effect of the Austrian Court’s local assessment. It needs to be recalled that a fundamental tenet of L’Oréal is the principle that the mere accessibility of a website is not enough a reason to conclude that a jurisdiction is being targeted, and it is for national courts to make the assessment. With the exception of the ECtHR that applies to date one of the most expansive jurisdictional approaches (Perrin, 2005), international policy-makers (such as for instance the UN Special Rapporteur on Freedom of Opinion and Expression, the OSCE Representative on Freedom of the Media, the OAS Special Rapporteur on Freedom of Expression and the ACHPR Special Rapporteur on Freedom of Expression and Access to Information’s Joint Declaration on Freedom of Expression and the Internet of 2011, among others) and most courts favour a different approach inspired to judicial self-restrain, and put emphasis on a ‘real and substantial connection’ to justify jurisdiction over Internet content.

When personality rights are at stake, the recent CJEU decision in Bolagsupplysningen of 2017 (incidentally, derogating from the more established CJEU decisions in Shevill, 1995, and eDate, 2011) suggested that, when incorrect online information infringes personality rights, applications for rectification and/or removal are to be considered single and indivisible, and a court with jurisdiction can rule on the entirety of an application.

However, this precedent (already controversial in its own right) seems unfit to apply to this case. Bolagsupplysningen builds at least on the same rationale as in the other decisions of the CJEU mentioned above, that expansive jurisdiction can be justified by the necessity to guarantee the protection of citizens’ fundamental rights and not to see them frustrated by a scattered territorial application. In the case of political speech, where limitations need to be justified by an overriding public interest, such as typically public safety, the connecting element the Court looks for becomes immediately less apparent, as it cannot be assumed that the same speech would prove equally inflammatory in different places under different social and political circumstances. In other words, public order better lends itself to territorially sensitive protection.     

This taps into the assessment of the necessity and proportionality of the measure that decision-makers need to make before removing content. As a matter of principle, the geographical scope of limitations is part of the least restrictive means test: the ECtHR for instance in Christians Against Fascism and Racism (1980) considered that even when security considerations would outweigh the disadvantage of suppressing speech and thus justify the issue of a ban, said ban would still need a ‘narrow circumscription of its scope in terms of territorial application’ to limit its negative side effects.  Similarly, the 2010 OSCE/ODIHR – Venice Commission Guidelines on Freedom of Peaceful Assembly (as quoted by the ECtHR in Lashmankin v. Russia of 2017) consider that ‘[b]lanket restrictions such as a ban on assemblies in specified locations are in principle problematic since they are not in line with the principle of proportionality which requires that the least intrusive means’, and thus restrictions on locations of public assemblies need to be weighed against the specific circumstances of each case. Translated in the context of digital communications, the principle requires that the territorial scope of content removal orders is narrowly circumscribed and strictly proportionate to the interest protected. An injunction to remove commentary on national politics worldwide, as a result, seems unlikely to be considered the least intrusive means.

Conclusions

The decision of the CJEU has the potential to change the landscape of intermediary responsibility and the boundaries of lawful speech as we know them. By being asked to remove content that is identical or analogous, intermediaries will be making, in all other cases other than removing mere copies of posts that have already been found illegal, active determinations on the legality of third-party content. While the re-upload of illegal content is an issue of growing importance that needs to be addressed, consideration needs to be paid as to whether this would be an appropriate measure, as solutions borrowed from other fields like copyright protection can sit at odds with the specificities of content regulation and infringe on European and international standards for the protection of freedom of expression online.

Similarly, by granting an extraterritorial injunction in this case the Court would follow a stream that has been emerging in the last few years in privacy and data protection. Thanks to extraterritorial reach, the GDPR is rapidly becoming a global regulatory benchmark for the processing of personal data, which arguably benefits European Union citizens and protects their relevant fundamental rights. The same could not be true if this rationale would be applied to standards of legitimate political speech. It is questionable whether the EU (or any other jurisdiction) bears any interest in setting a global regulatory benchmark for content regulation. By restricting the accessibility of content beyond the national boundaries where the original dispute took place, it would restrict other citizens’ right to receive information without granting any substantive benefit, such as protecting public security, to the citizens of the first state.

Barnard & Peers: chapter 9

Photo credit: Slate magazine

---

[1] L. Karamanidou (2016) ‘Violence against migrants in Greece: beyond the Golden Dawn’, Ethnic and Racial Studies, 39:11, 2002-2021; D. Skleparis (2016) ‘(In)securitization and illiberal practices on the fringe of the EU’, European Security, 25:1, 92-111.

Crushing terrorism online – or curtailing free speech? The proposed EU Regulation on online terrorist content

Professor Lorna Woods, University of Essex

On 12th September 2018, the Commission published a proposal for a regulation (COM(2018) 640 final) aiming to require Member States to require certain internet intermediaries to take proactive if not pre-emptive action against terrorist content on line as well as to ensure that state actors have the necessary capacity to take action against such illegal content. It is described as “[a] contribution from the European Commission to the Leaders’ meeting in Salzburg on 19-20 September 2018”. The proposal is a development from existing voluntary frameworks and partnerships, for example the EU Internet Forum, and the non-binding Commission Recommendation on measures to effectively tackle illegal content online ((C(2018)1177 final), 1st March 2018) and its earlier Communication on tackling illegal content online (COM(2017) 555 final). In moving from non-binding to legislative form, the Commission is stepping up action against such content; this move may also be seen as part of a general tightening of requirements for Internet intermediaries which can also be seen in the video-sharing platform provisions in the revised Audiovisual Media Services Directive and in the proposals regarding copyright. Since the proposal has an “internal market” legal base, it would apply to all Member States.

The Proposal

Article 1 of the proposed Regulation sets out its subject matter, including its geographic scope.  The scope of the proposed regulation is directed to certain service providers, “hosting service provider” in respect of specified content “illegal terrorist content”.  Terms are defined in Article 2. Article 2(1) defines “hosting service provider” (HSP) as “a provider of information society services consisting in the storage of information provided by and at the request of the content provider and in making the information stored available to third parties”. The definition of illegal terrorist content found in Article 2(5) is one (or more) of the following types of information:

(a) inciting or advocating, including by glorifying, the commission of terrorist offences, thereby causing a danger that such acts be committed;

(b) encouraging the contribution to terrorist offences;

(c) promoting the activities of a terrorist group, in particular by encouraging the participation in or support to a terrorist group within the meaning of Article 2(3) of Directive (EU) 2017/541

(d) instructing on methods or techniques for the purpose of committing terrorist offences.

The format does not matter: thus terrorist content can be found in text, images, sound recordings and videos.

Article 3 specifies the obligations of the HSPs. In addition to a specific obligation to prohibit terrorist content in their terms and conditions, HSPs are obliged to take appropriate, reasonable and   proportionate actions against terrorist content, though those actions must take into account fundamental rights, specifically freedom of expression.

Article 4 introduces the idea of a removal order, and requires that the competent authorities of the Member States are empowered to issue such orders; requirements relating to removal orders are set out in Article 4(3).  It does not seem that the issuing of such orders require judicial authorization, though the Regulation does envisage mechanisms for HSPs or the “content provider” to ask for reasons; HSPs may also notify issuing authorities when the HSP views the order as defective (on the basis set out in Article 4(8)), or to notify the issuing authority of force majeure. Article 4(2) states:

Hosting service providers shall remove terrorist content or disable access to it within one hour from receipt of the removal order.

The regulation also envisages referral orders; these do not necessitate the removal of content, nor – unlike the position for removal orders – does it specify deadlines for action. On receipt of a referral order, a HSP should assess the notified content for compatibility with its own terms and conditions. It is obliged to have in place a system for carrying out such assessments. There is also an obligation in Article 6 for HSPs in appropriate circumstances to take (unspecified) effective and proportionate proactive measures and must report upon these measures. Article 6 also envisages the possibility that competent authorities may – in certain circumstances – require a hosting service provider to take specified action.

Article 7 requires hosting service providers to preserve data for certain periods.  The hosting service provider is also required to provide transparency reports as well as to operate within certain safeguards specified in Section III, including transparency reporting, human oversight of decisions, complaints mechanisms and information to content providers – these are important safeguards to ensure that content is not removed erronously.  Section IV deals with cooperation between the relevant authorities and with the HSPs.  Cooperation with European bodies (e.g. Europol) is also envisaged.  As part of this, HSPs are to establish points of contact.

The Regulation catches services based in the EU but also those outside it which provide services in the EU (with jurisdiction in relation to Article 6 (proactive measures), 18 (penalties) and 21 (monitoring) going to the Member State in which the provider has its main establishment) and should designate a legal representative. The Member State in which the representative is based has jurisdiction (for the purposes of Articles 6, 18 and 21). Failure so to designate means that all Member States would have jurisdiction.  Note that as the legal form of the proposal is a Regulation, national implementing measures would not be required more generally.

Member States are required to designate competent authorities for the purposes of the regulation, and also to ensure that penalties are available in relation to specified articles such penalties to be effective, proportionate and dissuasive.  The Regulation also envisages a monitoring programme in respect of action taken by the authorities and the HSPs.  Member States are to ensure that their competent authorities have the necessary capacity to tackle terrorist content online.

Preliminary Comments

The proposal is in addition to the Terrorism Directive, the implementation date for which is September 2018.  That directive includes provisions requiring the blocking and removal of content; is the assumption that – even before they are require legally to be in place – these provisions are being seen as ineffective.

This is also another example of what seems to be a change in attitude towards intermediaries, particularly those platforms that host third party content.  Rather than the approach from the early 2000s – exemplified in the e-Commerce Directive safe harbour provisions – that these providers are and to some extent should be expected to be content-neutral, it now seems that they are being treated as a policy tool for reaching content viewed as problematic.  From the definition in the Regulation, it seems that some of the HSPs could have – provided they were neutral -fallen within the terms of Article 14 e-Commerce Directive: they are information society service providers that provide hosting services.  The main body of the proposed regulation does not deal with the priority of the respective laws but in terms of the impact on HSPs, the recitals claim

“any measures taken by the hosting service provider in compliance with this Regulation, including any proactive measures, should not in themselves lead to that service provider losing the benefit of the liability exemption provided for in that provision.  This Regulation leaves unaffected the powers of national authorities and courts to establish liability of hosting service providers in specific cases where the conditions under Article 14 of Directive 2000/31/EC for liability exemption are not met”.

This reading in of what is effectively a good Samaritan saving clause follows the approach that the Commission had taken with regard to its recommendation – albeit in that instance without any judicial or legislative backing.  Here it seems that the recitals of one instrument (the Regulation) are being deployed to interpret another (the e-Commerce Directive). 

The recitals here also specify that although Article 3 puts HSPs under a duty of care to take proactive measures, this should not constitute ‘general monitoring’; such general monitoring is precluded according to Article 15 e-Commerce Directive. How this boundary is to be drawn remains to be seen. Especially as the regulation envisages prevention of uploads as well as swift take-downs. Further, recital 19 also recognises that

“[c]onsidering the particularly grave risks associated with the dissemination of terrorist content, the decisions adopted by the competent authorities on the basis of this Regulation could derogate from the approach established in Article 15(1) of Directive 2000/31/EC, as regards certain specific, targeted measures, the adoption of which is necessary for overriding public security reasons”.

This is a new departure in the interpretation of Article 15 e-Commerce Directive.

The Commission press release suggests the following could be caught: social media platforms, video streaming services, video, image and audio sharing services, file sharing and other cloud services, websites where users can make comments or post reviews. There is a limitation in that the content hosted should be made available to third parties. Does this mean that if no one other than the content provider can access the content, the provider is not an HSP?  This boundary might prove difficult in practice.  The test does not seem to be one of public display so services where users who are content providers can choose to let others have access (even without the knowledge of the host) might fall within the definition. What would be the position of a webmail service where a user shared his or her credentials so that others within that closed circle could access the information? Note that the Commission is also envisaging services whose primary purpose is not hosting but which allows user generated content– e.g. a news website or even Amazon – also fall within the definition. 

The scope of HSP is broad and may to some extent overlap with that of video-sharing platforms or even audiovisual media service providers for the purposes of the Audiovisual Media Services Directive (AVMSD).  Priorities and conflicts will need to be ironed out in that respect. The second element of this broadness is that the HSP provisions are not just applying to the big companies, the ones to some extent already cooperating with the Commission, but also to small companies. In the view of the Commission terrorist content may be spread just as much by small platforms as large.  Similar to the approach in the AVMSD, the Commission claims that the regulatory burden will be proportionate as the proportionality with mean the level of risk as well as the economic capabilities would be taken into account. 

In line with the approach in other recent legislation (e.g. GDPR, video-sharing platforms provisions in AVMSD) the proposal has an extraterritorial dimension. HSPs would be caught if they provide a service in the EU.  The recitals clarify that “the mere accessibility of a service provider’s website or of an email address and of other contact details in one or more Member States taken in isolation should not be a sufficient condition for the application of this Regulation” [rec 10]; instead a substantial connection is required [rec 11]. Whether this will have a black out effect similar to the GDPR remains to be seen; it may depend on whether the operator is aware enough of the law; how central the hosting element is and how large a part of its operations the EU market is.

While criminal law, in principle, is a matter for Member States, the definition of terrorist content relies on a European definition – though whether this definition is ideal is questionable.  For companies that operate across borders, this is presumably something of a relief (and as noted above, the proposal is based on Article 114 TFEU, the internal market harmonisation power).  The Commission also envisages this a mechanisms limiting the possible scope of the obligations – only material that falls within the EU definition falls within the scope of this obligation – thereby minimising impact on freedom of expression (Proposal p. 8).  Whether national standards will consequently be precluded is a different question.  Note that the provisions in the AVMSD that focus on video sharing platforms were originally envisaged as maximum harmonisation but, as a result of amendments from the Council, retuned to minimum harmonisation (the Council amendments also introduced provisions on terrorist content into the AVMSD based on the same definition).

The removal notice is a novelty aimed at addressing differential approaches in the Member States in this regard (an on-going problem within the safe harbour provisions of the e-Commerce Directive), but also to ensure that such take down requests are enforceable.  Note, however, that it is up to each Member State to specify the competent authorities, which may give rise to differences between the Member States, perhaps also indicating differences in approach.  The startling point is probably the very short timescale: 1 hour (a complete contrast to the timing for example specified in the UK’s Terrorism Act 2006).  The removal notices have been a source of concern.  This is not very long which will mean that - especially with non-domestic providers and taking into account time differences - HSPs will need to think how to man such a requirement (unless the HSPs plan to automate their responses to notices), especially if the HSP hopes to challenge ‘unsatisfactory’ notices (Art 4(8)). 

Given the size of the penalties in view, industry commentators have suggested that all reported content will be taken down.  This is certainly would be a concern in relation to situations where the HSPs had to identify terrorist content (ie ascertain not just that it was in a certain location but also that it met the legal criteria) themselves.  Is it not the case that this criticism is fully appropriate here.  Here, HSPs are not having to decide whether or not the relevant content is terrorist or not- the notice will make that choice for them.  Further, the notice is made not by private companies with a profit agenda but instead by public authorities (presumably) orientated to the public good and with some experience in the topic as well as in legal safeguards.  Furthermore, the authority must include reasons. Indeed, the Commission is of the view that referrals are limited to the competent authorities which will have to explain their decisions ensures the proportionality of such notices (Proposal p. 8). Nonetheless, a one hour time frame is a very short period of time.

Another ambiguity arises in the context of referral notices. It seems that the objective here is to put the existing voluntary arrangements on a statutory footing but with no obligation on the HSP to take the content down within a specified period. Rather the HSP is to assess whether the content referred is compatible with the HSPs terms of service (not whether the content is illegal terrorist content).  Note this is a different from the situation where the HSP discovers the content itself and there has been no official view as to whether the content falls within the definition of terrorist content or not. This seems rather devoid of purpose: relevant authorities have either decided that the content is a problem (in which case the removal notice seems preferable as the decision is made by competent authorities not private companies) or the notice refers to content which is not quite bad enough to fall with the content prohibited by the regulation but the relevant authorities would still like it down, with the responsibility for that decision being pushed on to the HSP. Such an approach seems undesirable.

Article 6 requires HSPs to take effective proactive measures.  These are not specified in the Regulation, and may therefore allow the HSPs some leeway to take measures that seem appropriate in the light of each HSP’s own service and priorities, though it seems here that there may also be concerns about the HSPs’ interpretation of relevant terrorist content.  It is perhaps here that criticisms about the privatisation of the fight against terror comes to the fore.  Note, however that Article 6(4) allows a designated authority to impose measures specified by the authority on the HSP.  Given that this is dealt with at the national level, some fragmentation across the EU may arise; there seems to be no cooperation mechanism or EU coordination of responses under Article 6(4).

There is also the question of freedom of expression. Clearly state mandated removal of content should be limited, but it is the intention that HSPs have no freedom to remove objectionable content for other reasons. At some points, the recitals suggest precisely this: “hosting service providers should act with due diligence and implement safeguards, including notably human oversight and verifications, where appropriate, to avoid any unintended and erroneous decision leading to removal of content that is not terrorist content” [rec 17]. Presumably the intention is that HSPs should take steps to avoid mistakenly considering content to be terrorist.  They clearly are under obligations to take other forms of content down, e.g. child pornography and hate speech. 

More questionable is the position with regard other types of content: the controversial and the objectionable, for example.  As private entities human rights obligations do not bite on them in the same way as they do with regards to States, so there may be questions about the extent to which a content provider can claim freedom of expression against an unwilling HSP (e.g. for Mastodon, the different instances have different community standards set up by that community - should those communities not be entitled to enforce those standards (providing that they are not themselves illegal)?).  There may moreover be differences between the various Member States as to how such human rights have horizontal effect and the deference given to contractual autonomy.  With regard to the video sharing platforms, it seems that room is given to the platforms to enforce higher standards if they so choose; there is not such explicit provision here.

A final point to note is the size of the penalties that are proposed.  The proposal implicitly distinguished between one-off failings and a ‘systematic failure to comply with obligations’.  In the latter cases, penalties of up to 4% of global turnover- in this there are similarities to the scale of penalties under the GDPR.  This seems to be developing into a standard approach in this sector.

Barnard & Peers: chapter 25, chapter 9

JHA4: chapter II:5

Photo credit: Europol