This site would like to set some non-essential temporary cookies. Some cookies we use are essential to make our site work.
Others such as Google Analytics help us to improve the site or provide additional but non-essential features to you.
No behavioural or tracking cookies are used.
To change your consent settings, read about the cookies we set and your privacy, please see our Privacy Policy

Liability: Developments in intermediary liability: the cases so far

Online intermediary liability has been debated in the courts since the internet came into being. Following a succession of recent court cases little clarity has resulted on the circumstances surrounding when an intermediary will and will not be liable for unlawful content posted on its networks. Robert Blamires and David Naylor, of Field Fisher Waterhouse in London, discuss the most important cases involving the hosting platforms provided by Google and Amazon, and how far the 'Hosting Defence' protects intermediaries from liability as a publisher of defamatory content.


The question of the extent to which an online intermediary, such as an internet service provider, a search engine, or a social network, should be liable for content transmitted using their services, has been vigorously debated for as long as the internet has been in existence. In fact, the broader question of whether an intermediary provider of technology should be liable for third party infringing / illegal use pre-dates the internet. Some readers will remember twin cassette deck machines, and may also remember the case brought by rights-holders against Amstrad in relation to the marketing and supply of such machines1. In that case the fact that Amstrad's machines could be used for both infringing and non-infringing purposes and that Amstrad had no control over how the machines were used, once purchased, convinced the House of Lords that Amstrad was not authorising users to infringe copyright. Similarly, in the US, a few years earlier, the Supreme Court determined that Sony was not liable for copyright infringements committed by purchasers of its Betamax video tape recorder, on the grounds that the device had substantial non-infringing uses.

This debate has continued over the years since the Betamax and Amstrad cases, with claims being brought across the globe, involving intermediary technology, such as: peer-to-peer file sharing services Napster, Grokster and Kazaa; ISPs Yahoo!, Demon Internet, BT, Tele2, and Tiscali / Scarlet; eBay, the online marketplace; and Newzbin, the indexing service enabling users to locate and compile copyright-infringing content.

In the online space, there have, for a number of years, been legislative protections in place to protect some online intermediaries, in some circumstances, when their users transmit infringing content using their services. In the UK, these are found in the Electronic Commerce (EC Directive) Regulations 2002 (the 'E-Commerce Regulations'), which are based on the E-Commerce Directive2 (and therefore legislation similar to the E-Commerce Regulations is in place in Member States across Europe).

Recent Developments

In the last year or so there have been a number of high-profile cases concerning the liability of certain intermediaries for content posted using their services. Two such cases in the UK, concerned Google's liability for comments posted by third party users of (which is owned and operated by Google); Davison v Habeeb and others3 (the 'Davison case') and Tamiz v Google4 (the 'Tamiz case'), while a third concerned Amazon's liability for content posted on review sections of its website (the 'Amazon case')5.

Both Google cases involved complaints that content posted on was defamatory, and involved applications by Google to have UK orders for service on Google, Inc outside of the jurisdiction (ie in the USA), set aside and, in both cases, Google was successful. However, the reasoning behind the two cases is noticeably different. In the Tamiz case, Google prevailed on the 'in principle' arguments that its responsibility for did not, for the purposes of the English law of defamation, make it a publisher6, and that it was not authorising publication of defamatory material. In contrast, in the Davison case, Google prevailed because the court found that the claimant (Ms Davison) had failed to establish that there was a real and substantial tort within the jurisdiction, but the judge went on to hold that Google's responsibility for did make it a publisher of the allegedly infringing material.

However, despite the apparent conflict on whether Google was a publisher in relation to, in both cases the court held that Google had a defence under regulation 19 of the E-Commerce Regulations (the 'Hosting Defence'). Similarly, in the Amazon case HHJ Moloney held that the Hosting Defence was 'bound to succeed in relation to Amazon's liability' for the allegedly defamatory material.

This article examines the recent cases, and the outcomes on the applicability of the Hosting Defence, and considers the risk to internet intermediaries of being put on notice of infringing content on their networks.

The Davison case

Ms Davison alleged that a number of articles written by Mr Peter Eyre, and published on websites including on a blog hosted by, were and, at some point after notification (of the material complained of ), became liable for the continued publication of that material.

The Tamiz case

The Tamiz case involved a similar set of facts to the Davison case. Mr Tamiz alleged statements made in a blog hosted by were defamatory, he notified Google and, whilst Google did not itself remove the statements, it had contacted the author in question, who did do so. Mr Tamiz obtained an order to serve Google out of the jurisdiction and Google applied to have it set aside.

The judge (Eady J) having concluded that a real and substantial tort had been committed within the jurisdiction, considered whether Google was a publisher, concluding that Google's role with respect to is passive, and comparable to an owner of a wall which has been festooned overnight with defamatory graffiti. Eady J commented that 'it is no doubt true that the owner of that wall could acquire scaffolding and have it all deleted by whitewash. That is not necessarily to say, however that the unfortunate owner must, unless and until this has been accomplished be classified as a publisher.' He was also persuaded by the fact that Google, whether or not notified of a complainant's objection to content, was not required to take any technical step to continue the accessibility of the offending material, ie it did not have to positively act for the content to continue being available. As a result, Eady J concluded that Google should not be regarded as a publisher or even as one who authorises publication under the established principles of the common law.

The Amazon case

The Amazon case concerned (amongst other things) an application by Amazon to have a claim against it for allegedly libellous content, posted on a review section of its website, struck out on the grounds that it was bound to be able to rely on the Hosting Defence (and/or the defence under section 1 of the Defamation Act). The judge considered in detail both Amazon's pre- and post-publication processes for handling potentially illegal content, namely its guidelines, ability to blacklist offending users, automatic filtering, 'Report Abuse' function, Notice and Takedown Procedure and General Conditions of Use and Sale, concluding that, for the purposes of sections 1(2) and 1(3) of the Defamation Act, Amazon was 'plainly not' a commercial publisher and 'clear[ly] ... not' an editor.

The Hosting Defence

In both Google cases, the court (albeit obiter in the Tamiz case) held that Google could rely on the Hosting Defence. In the Tamiz case, Google had received a complaint which had simply alleged that content on was defamatory, with no explanation as to why, and, for Google to have been fixed with actual knowledge of the unlawfulness of that content, it would need to have known something of the strengths and weaknesses of the available defences. In the Davison case, Google had received conflicting claims from the claimant and the author of the allegedly defamatory articles which 'it was in no position to adjudicate' (although the court would have been willing to reach a different conclusion in different circumstances, such as if the complaint had been 'sufficiently precise and well substantiated').

Similarly, in the Amazon case, the court held that Amazon could rely on the Hosting Defence, on the grounds that: (i) there was 'no reason to suppose that anyone in Amazon was actually aware of [the allegedly defamatory] postings, let alone of their possible unlawfulness, prior to [the claimant's] complaints'; and (ii) following such complaints: Amazon could not have discriminated between legitimate 'comment in strong terms' and 'less obviously legitimate areas' because, although the claimant had identified the user in question and the topic of the comments, he had not specified which particular postings, or parts of posting, of which he complained. Further, even if the complainant had done so, the Hosting Defence 'demands more from a complainant, if the defence is to be overcome. He must disclose facts or circumstances making it apparent that the postings were unlawful.' Amazon's pro forma for complaints recognises this and is structured to invite such disclosure (requiring, amongst other things) a signed document forming a basis for taking action which might be harmful to the other party) but the claimant had failed both to use the pro forma and to address this disclosure requirement at all, making it impossible for Amazon to investigate or adjudicate on the merits of the claimant's case.

The theme common to all three judgments, in relation to the Hosting Defence, is that giving an information society service provider notice of allegedly unlawful content is not, on its own, necessarily enough to give the provider knowledge of unlawful activity or information, so as to lose the benefit of the Hosting Defence. Some further knowledge of the unlawfulness of the activity or information is needed.

Government reform

Against this backdrop, the Government has recently acknowledged that, in the context of the draft Defamation Bill, the law has not kept pace with the development of modern communication culture7. Whilst the Bill is still subject to further consultation, it appears that the Government favours a greater degree of protection for internet intermediaries; it includes a statutory defence to a defamation action in respect of a statement posted on the website, for a website operator that can show that it did not post the statement in question, as well as provisions which: will protect internet archives; limit forum shopping against defendants based outside the EU or EEA (eg a US website accessible in England and Wales); and bar defamation actions against secondary publishers (eg a website) unless it is not reasonably practicable to proceed against the author, editor or commercial publisher.


The cases discussed in this article may give some comfort to intermediaries, such as Google and Amazon, that they will not automatically be exposed (by losing the benefit of the Hosting Defence) in relation to every notice / communication they receive about allegedly illegal or infringing content on their networks.

However, while there will always be examples of clear-cut infringements / illegality, the 'something more' message leaves little room for certainty in the grey areas. How 'precise and well substantiated' would a complaint have to be for the intermediary to be on notice? Can providers relax, safe in the knowledge that nothing but a notice resembling a court submission, listing each and every breach, together with statutory / common law references and full substantiating evidence, will put them on notice? It is, of course, unlikely that many individual claimants will be in a position to provide this level and quality of information. In any case, how is an intermediary supposed, even with some facts enabling it to attempt to do so, to take the position of adjudicator and attempt to judge, in relation to any complaint (let alone potentially thousands) the merits of the case? Intermediaries are of course likely to be alive to these points and the potential uncertainty they create.

In any event, while the judicial and legislative tide may currently be flowing a little towards intermediaries, and a well-structured notification procedure may assist diligent operators, the dividing lines between when an intermediary will and will not be liable for content on its networks, and when an intermediary is and is not on notice of unlawful content, are still by no means clear.

Robert Blamires

Senior Associate

David Naylor


Field Fisher Waterhouse

1. CBS Songs Limited and others v Amstrad Consumer Electronics Plc and others [1988] AC 1013.

2. Directive 2000/31/EC of the European Parliament and of the Council of 8 June 2000 on certain legal aspects of information society services, in particular electronic commerce, in the Internal Market.

3. [2011] EWHC 3031 (QB).

4. [2012] EWHC 449 (QB).

5. [2012] EWHC B3 (QB).

6. And therefore has a defence to defamation proceedings under section 1 of the Defamation Act 1996 (the 'Defamation Act').

7. Response to the Report of the Joint Committee on the Draft Defamation Bill.