European Court of Human Rights holds news portal liable for user comments

Dick from the Internet

The European Court of Human Rights (ECtHR) has dealt a blow to existing intermediary liability rules in Europe in the case of Delfi v Estonia.  The decision of the Grand Chamber reaffirms an earlier decision by the first section of the court. For years we have had a system of limitation of intermediary liability based on the E-Commerce Directive, a regime that is now under fire.

Facts of the case

Delfi is one of the largest Internet portals in Estonia, and it offers content and news items in Estonian and Russian. Like most news websites in the world, published content has a comment section at the bottom of the article. The website has in place a flagging system for comments inciting hatred, which could culminate in the expeditious take-down of a comment. Similarly, the system deletes comments containing obscene words, and defamatory statements can be immediately removed upon request from the offended party.

Delfi has made publicly available terms of use where it specifies the type of conduct that is not permitted on the comment section; including the posting of messages containing threats, obscenity, insults, incitations to violence, and incitations to commit illegal acts, among other offences.

So far, so good. This is normal and standard procedure.

However, Delfi had built a reputation for being a place where disrespectful comment was repeatedly posted, at least according to the government. On September 2005, a local newspaper posted an open letter to the Minister of Justice lamenting the state of abuse online, and singling out Delfi as “a source of brutal and arrogant mockery”.

On January 2006, Delfi posted an article about a ferry company destroying ice roads on purpose. Ice roads are public open roads over the frozen sea in the winter, so the implication was that the ferry company wanted to get rid of these to protect its profits. The article named ‘L.’ as the major shareholder of the company at the time. Needless to say, the idea of a company destroying open and free roads for profit was met with anger, and in a short time the article had accumulated 185 comments, many of them insulting and even threatening L. with violence. Almost two months after publication, L. sent a request to Delfi to remove the comments, and asking for 500,000 EEK in damages. Delfi removed the comments the same day, but refused to pay. However, it is important to note that there was indeed a lag between publication and removal of comments of six weeks. ETA: but it must also be pointed out that Delfi did remove the user-generated content as soon as it received the complaint. See comments.

L. brought a civil suit against Delfi, and the lower court dismissed the claim based on the Estonian transposition of the E-commerce Directive using the standard rules for intermediaries, namely that Delfi was not the publisher of the comments, and had removed them promptly as required. It all should have ended there, but L. appealed successfully, as the court of appeals considered that the lower court was wrong in dismissing the case, and sent it back. This is where the case takes a weird legal turn, as the lower court decided not to use the E-Commerce Directive, but its Obligations legislation. The posting of the terms and conditions were considered to create an obligation on the parties to behave appropriately, and the County Court considered that the existing terms did not protect personality rights adequately. While the originating article was balanced, the comments were insulting, so Delfi were told to pay 5,000 EEK. Delfi appealed, but the court of appeal upheld the County Court’s decision. Delfi appealed again to the Estonian Supreme Court, which also dismissed the appeal.

Delfi decided to go all the way to the European Court of Human Rights, alleging that holding it liable for the comments posted by the readers of its Internet news portal infringed its freedom of expression rights under Article 10 of the European Convention of Human Rights (ECHR). As stated, the first section of the court decided against Delfi.


From the start, it feels like bringing the case all the way to the ECtHR was a mistake on the part of Delfi, particularly because it based the case on the very weak argument that freedom of speech had been violated. Having said that, it is really strange that the case got this far with so many courts deciding to ignore for the most part the E-Commerce Directive and the very rich case law from the Court of Justice of the European Union (CJEU) about the exemption of liability given to intermediaries.

Regardless, the majority opinion makes for some truly depressing reading.  Right off the bat, the concurring opinion by Judges Raimondi, Karakas, De Gaetano and Kjølbro states categorically that there is no violation of freedom of speech. Then they base their opinion strongly on the decision by the Estonian Supreme Court. The judges say:

“4. There are two possible readings of the Supreme Court’s judgment: (1) Delfi was liable as it did not “prevent” the unlawful comments from being published, and its liability was aggravated by the fact that it did not subsequently “remove” the comments; or (2) Delfi did not “prevent” the unlawful comments from being published, and as it did not subsequently “remove” the comments without delay, it was liable for them.
.  The Court has decided to read the Supreme Court’s judgment in the second sense, thereby avoiding the difficult question of the possible liability of a news portal for not having “prevented” unlawful user-generated comments from being published. However, had the Court read the Supreme Court’s judgment in the first sense, the outcome of the case might have been different.”

Things could have been much worse! Imagine if the ECtHR had decided that news organisations were under an obligation to pre-censor content! Well, we get the next worst thing, which is that the judges opined that Delfi had erred because it had not removed the comments immediately.

In my view, the killer paragraph in the opinion is this one:

“11. In our view, member States may hold a news portal, such as Delfi, liable for clearly unlawful comments such as insults, threats and hate speech by readers of its articles if the portal knew, or ought to have known, that such comments would be or had been published on the portal. Furthermore, member States may hold a news portal liable in such situations if it fails to act promptly when made aware of such comments published on the portal.

This sets a horrible new principle of intermediary liability that is completely at odds with the current system. The existing principle has been to extend an exemption of liability to services after they have been made aware of the damaging content, and Delfi did just that. The principle set above creates a horrendously open language that says that the portal “knew, or ought to have known, that such comments would be or had been published“. What does that even mean? It’s almost pre-crime, it puts an obligation on content services to guess when insulting comments may take place. The judges do not help when they explain what they mean:

12.  The assessment of whether the news portal knew or ought to have known that clearly unlawful comments may be or have been published on the portal may take into account all the relevant specific circumstances of the case, including the nature of the comments in question, the context of their publication, the subject matter of the article generating the comments, the nature of the news portal in question, the history of the portal, the number of comments generated by the article, the activity on the portal, and how long the comments have appeared on the portal.”

In my view, this may put a chilling effect on controversial topics, as it is clear that subject matter may be used to decide if a portal ought to have known that infringing comments might occur. Let’s not publish on controversial subjects! Let’s only publish articles within a narrow remit of safe subjects! Let’s continue with the buzzfeedification of content and make sure that we only publish inane lists and cat pics!

As if things were not bad enough, we get a reactionary concurring opinion by judge Zupančič, which should go down in infamy in the annals of Internet Law. He says:

Also, in my opinion, it is completely unacceptable that an Internet portal or any other kind of mass media should be permitted to publish any kind of anonymous comments. We seem to have forgotten that “letters to the editor”, not so long ago, were double-checked as to the identity of the author before they were ever deemed publishable. The Government argued (see paragraph 90 of the judgment) that the biggest international news portals did not allow anonymous (that is, unregistered) comments and referred to an opinion that there was a trend away from anonymity. At the same time, anonymous comments tended to be more insulting than the comments by persons who had registered, and harsh comments attracted more readers. The Government argued that Delfi had been notorious for exactly this reason.

Take a minute to digest that paragraph and absorb its awfulness. There are people posting anonymously online! This is a violation of the sacred protection of personality rights that goes back to the Romans! Will anyone rid me of these pesky anonymous electronic peasants?

Dissenting opinion

But not everything is lost. A strong dissenting opinion by judges Sajó and Tsotsoria offers us much-needed relief in a sea of bad news. The judges very clearly notice that the principle of active knowledge may have nefarious consequences. They say:

” In this judgment the Court has approved a liability system that imposes a requirement of constructive knowledge on active Internet intermediaries (that is, hosts who provide their own content and open their intermediary services for third parties to comment on that content). We find the potential consequences of this standard troubling. The consequences are easy to foresee. For the sake of preventing defamation of all kinds, and perhaps all “illegal” activities, all comments will have to be monitored from the moment they are posted. As a consequence, active intermediaries and blog operators will have considerable incentives to discontinue offering a comments feature, and the fear of liability may lead to additional self-censorship by operators. This is an invitation to self-censorship at its worst.”

I agree with every word, this is a horrible principle that may create an incentive for commercial news services to curb any sort of online discussion. We can hope that this is one of those situations where the dissenting opinion survives and becomes the prevalent principle cited by national courts.

Conflict with CJEU case law

Something that has been hinted at already is that this ruling is in direct contradiction with the E-Commerce Directive as it creates the principle of constructive knowledge that is not present in the current regime. Furthermore, Art 15 of the Directive specifically states that intermediaries are under no obligation to monitor content that they transmit or host. Moreover, the decision is also in conflict with various CJEU cases, including Google France and Google (C-236/08 to C-238/08), L’Oréal (C-324/09), SABAM (C-360/10), and Scarlet Extended (C-70/10).

Interestingly, the ECtHR does cite these, and two more, Google Spain (C-131/12 of RTBF fame), and Papasavvas (C-291/13). I say interestingly because this may be the first time that Google Spain is being used in the context of intermediary liability, and may not be the last (but I digress). I also think that citing Papasavvas is interesting because it is irrelevant in this situation, as that was a case where a newspaper published potentially defamatory statements in its printed edition, which then were published online. The newspaper was trying to argue that the online publication somehow meant that it could use the exception of limitation of liability, but the court clearly failed to do so.

So, what happens in case of a conflict between the CJEU and the ECtHR? I am not an EU lawyer (and I’m happy to be corrected if the following is inaccurate), but from what I have read this is a complicated issue. The ECtHR is not part of the EU system, in fact there are signatory countries to the ECHR that are not in the EU, including Russia and Turkey. However, the ECtHR rulings are binding only to the countries concerned in the decision, and they do not change the law, although countries have modified legislation to respond to rulings.

The immediate effect is that the decision will be followed in Estonia, so bad news for Delfi. It is possible that the decision would be cited by national courts, which is what we do not want.

As with regards to the CJEU, the ECtHR can hold countries accountable for human rights violations, even if they are following EU law. However, things could have been worse because we are in the middle of the process of accession of the CJEU into the ECHR. However, last year the CJEU gave a negative opinion (Opinion 2/13) stating that at the moment, accession “is liable adversely to affect the specific characteristics and the autonomy of EU law”. If we had accession, then the Delfi decision would have been binding for the CJEU.

What next?

The worst case scenario is that newspapers and other commercial websites will not care about the intricacies of European procedure and they will decide to curb comments and maybe even removing them altogether.

The dystopian nightmare scenario is that people will take Zupančič’s opinion seriously and will remove anonymous commenting facilities altogether.

The best case scenario is that national courts may choose to ignore Delfi and continue to use the existing case law from the CJEU, but I have no basis on which to base this opinion other than a feeling that the CJEU will simply not look kindly on a decision that largely ignores its own practice.

The rainbows and unicorns scenario is that the a national court refers an intermediary liability case to the CJEU, and it cites Delfi directly only to nuke it to kingdom come.

May you live in interesting times.

Comments 10

  1. I do not really get this blog post.

    Namelt you write that:
    “Things could have been much worse! Imagine if the ECtHR had decided that news organisations were under an obligation to pre-censor content! Well, we get the next worst thing, which is that the judges opined that Delfi had erred because it had not removed the comments immediately.”

    Well… I was representing Delfi and thus know that Delfi removed the comments the same day as it got the notice to take them down. In legal terms – expeditiously. See para 19 in decision.

    You refer to 6 weeks as time that took Delfi to remove. This is unfortunately wrong – thus we actually got the “worse case cenario”, since the case is about a platform that expeditiously removed UGC.


    1. My wording might be confusing, I’ll try to rephrase. I meant to say that the content was visible for 6 weeks, but that Delfi adequately removed the content immediately after it was told.

      Does that make sense?

Leave a Reply