Hyperlinks are a fundamental part of the core fabric of the web. As the basic tool used to connect pieces of information together, it’s difficult to imagine how the Internet could function without them. 

Despite its critical nature, the role of hyperlinks has attracted the attention of those seeking to prevent particular kinds of information from being shared. One of the prime examples relates to copyright, and efforts to disrupt the dissemination of materials without authorisation. As part of this, the delivery mechanism of the hyperlink – as well as the infringing act itself – has come under fire.

Hyperlinks and Case Law

There is not a sizeable wealth of case law available that directly relates to the question of whether a hyperlink can constitute copyright infringement. As a result, discussions concerning potential liability often draw upon analogies taken from older cases – sometimes with judgements over a century old – in order to apply established legal principles to the uncertainties thrown up by technological evolution.

In the case of Hird v. Wood from 1984, the defendant was seated near to a sign on which defamatory messages were displayed. Despite not having created the sign, it was held that he incurred liability simply through the act of drawing attention to it. In the later case of Byrne v. Dean, Hird was referenced, the judge noting the following:

If defamatory matter is left on the wall of [a] premises by a person who has the power to remove the defamatory matter from the wall he can be said to have published the defamatory matter to the persons who read it.

Applied to our topic, it would appear that this principle would impose an obligation on the operators of websites (as well as their hosts) to remove hyperlinks that led to illegal material.

The Supreme Court in Crookes v. Newton did not completely accept the above analogy. Instead, they found the argument put forward by the respondent to be more persuasive: a comparison between hyperlinks and footnotes. In other words, both are ‘content neutral’, communicating the existence of something, but not necessarily commenting on the content. This view is one also supported by Tim Berners-Lee, credited with the creation of the World Wide Web:

The intention […] was that normal links should simply be references, with no implied meaning.

However, the court also recognised that the Internet is ‘a potentially powerful vehicle’ for defamation, and that the context itself was important in establishing possible liability. In the words of the court:

Individuals may attract liability for hyperlinking if the manner in which they have referred to content conveys defamatory meaning.

Analogy is one instrument that can be used when considering the relationship between the use of hyperlinks and the law; powerful, albeit imprecise. The issues involved in real life situations are more complex than can be addressed by analogy alone, and courts have often taken differing approaches in making their determinations.

One of the first cases to directly challenge the legality of the use of hyperlinks, and their potential to constitute an infringement of copyright concerned their use on the website of a Scottish newspaper named the ‘Shetland News’.

The links in question were published on the Shetland News website. They took the form of headlines copied from the site of a rival paper: the ‘Shetland Times’. By visiting the Shetland News website and clicking on these links, visitors were taken to the corresponding articles on the Shetland Times website. Confused yet? The way things were set up meant it was possible to completely avoid having to visit the Shetland Times’ homepage, and therefore missing out on its advertising. As a result, the Shetland News website was receiving ad-based income for providing direct links to articles that they themselves had not authored. 

Understandably, the Shetland Times weren’t too pleased about this, and succeeded in having the use of the links halted through the use of an interim interdict. The reasoning for the decision was that the links came in the form of headlines that had been copied verbatim from the other site, and so there was potential copyright infringement. Disappointingly from an academic point of view, the case was settled out of court, with no final judgement made on the actual liability arising from the use of the links.

One of the first significant cases in the US regarding the status of hyperlinks was that of Kelly v. Arriba Soft Corp. This concerned the display of thumbnail images from a professional photographer’s website in search results. The images were available both as resized thumbnails, and full sized previews, which is akin to the functionality provided by Google Images. Arriba was sued for copyright infringement, and the appeal judgement from the Ninth Circuit Court in San Francisco found that the thumbnails were protected under the doctrine of fair use. However, there was liability incurred for the displaying of the images in a new display window – a practice known as ‘in-line linking’. After an amicus brief filed by the EFF, the judgement was revised; the concerns about the in-line linking removed, with the fair use affirmation standing. 

In the case of Intellectual Reserve, Inc. v. Utah Lighthouse Ministry, Inc., the court found that hyperlinks pointing towards illegally distributed material could in of themselves be considered to be contributory copyright infringement. In this particular situation, the facts were complicated because the owner of the website in question had originally stored copies of the protected content on their own servers, before replacing them with hyperlinks to copies stored elsewhere. It was the context of these actions that was important – echoing the ratio decidendi in Hird.

One of the more commonly cited cases in this area is the infamous Grokster case, in which the Supreme Court introduced a new potential for liability: that of inducement. It was held that where technology is created for the intended or actual purpose of encouraging its users to breach copyright, then the creators themselves could be held liable for contributory copyright infringement. Ultimately, Grokster was shut down. Despite the concern by service providers over the precedent of this case, the facts were very particular to this situation, with the platform actively fostering the ‘blatant and overwhelming’ infringing activity of their users. It is extremely unlikely that the same definition would be applied to the majority of contemporary online intermediaries.

Hyperlinks, the DMCA, and Contributory Infringement

Online service providers often receive DMCA takedown notifications that target hyperlinks leading to allegedly infringing material, rather than material that resides on their servers. They are faced with an interesting quandary as a result: whether or not to remove the link.

On the one hand, hyperlinks generally do not constitute copyright infringement. However, it is the context that is the determining factor when considering potential for liability. A link created by a user to illegal material may well be infringing, but where does this leave the service provider?

Service providers are afforded safe harbor immunity from the infringing actions of their users, provided they ‘remove or disable access to’ material upon receipt of a valid DMCA takedown notification. However, it is unclear how this would apply in the case of hyperlinks. In our example, the infringing material itself is located on servers out-with their control, but there is still potentially infringing activity taking place on their platform. Would a host be liable for a failure to remove a hyperlink to material, where that hyperlink was found to be an infringing act, based on the context?

We can take some insight into how the decisions of future courts may fall in this scenario by considering the ‘server test’ discussed in the Perfect 10 cases. Here, the facts concerned the display of websites in Google’s Image search that were infringing upon the copyright interests of the plaintiff. It was found that Google was not liable for direct infringement on the basis that the material at issue did not reside on their servers, and was served up from another host through their use of framing, or ‘in-line linking’. With regards to contributory infringement specifically, the court held that there was no liability, as the infringing activity itself would still exist irrespective of whether or not Google Images existed. In other words, they were not found to be encouraging the copyright infringement.

In Flava, the defendants were operators of a ‘social bookmarking’ service called myVidster that allowed users to share videos from different locations around the web, which were then embedded on their platform, served up from the original locations. They were sued for contributory copyright infringement, based on the actions of users that were sharing clips that had been uploaded without authorisation.

The service provider had already received a number of takedown notifications regarding the material, and it was argued that they had not taken enough action as required to qualify for safe harbor protections. However, that is not the be all and end all. In the words of the court: ‘a non-infringer doesn’t need a safe harbor.’ The parties who uploaded the videos in the first place were the ones whose activity was infringing, and the question is whether myVidster had encouraged their infringing activity to an extent that constituted contributory infringement. The court did not find this to be the case, holding that myVidster was neither a direct or contributory infringer. In other words, they were too far removed from the infringing activity, 

My View

Irrespective of the potential for individuals to incur liability based on the context of hyperlinks which they create, the application of the DMCA should not extend to their removal.

Part of the criticism of the DMCA is that there is a substantial burden placed on copyright holders to track down and report instances of infringement across the web. Much like the mythological Hydra, as one falls, more spring up to take its place. Slaying the beast requires attacking the root of the problem; treating hyperlinks as valid subjects of takedown notifications is to mis-understand the task, and only serves in the creation of extra conceptual heads to pursue.

Rather than target hyperlinks, the focus of enforcement efforts should be on the actual source of the infringing material: the host. Take out the location pointed to by hyperlinks, and they are instantly rendered obsolete. This is not only a far more effective approach in tacking infringing activity, but one that also avoids creating extra and unnecessary work. Financially, this means less is paid to third party agents such as DMCA.com, whose revenue is based on successful takedown notifications.

The DMCA is already a blunt, and powerful tool. Abuse of the system is rife, and often deployed for the purposes of censoring legitimate expression, rather than to curb copyright infringement. To extend its remit to include the removal of hyperlinks is a dangerous step, that fundamentally alters our relationship with a core structural element of the web, and risks a (further) chilling effect on freedom of speech.

Despite recent case law seeming to support this principle, the judgements have been extremely dependent on the circumstances involved, and there has been no definitive authorities. As a result, it is up to online service providers to shape the approach to the issue, rejecting DMCA notification takedowns that concern hyperlinks. Policy decisions such as this create the normative frameworks that have the power to help ensure or hinder a free and open web, and it’s critical that tech companies lead the charge, rather than taking a minimum risk stance.