Searching for defamation law reform - Dr Daniel Joyce


This article was originally published in the Gazette of Law & Journalism, see:

We are grateful to the Gazette and Dr Joyce for allowing us to republish the article here.


Searching for defamation law reform

Who should bear liability for the publication of defamatory material via a search engine? UNSW media law academic Dr Daniel Joyce reviews the case law in Australia in light of UK law and an up-coming High Court consideration of the issues … There is an urgent need for clarity, and for statutory reform

By Dr Daniel Joyce*

Defamation law in Australia has traditionally cast a wide net in terms of liability.

In recent years increasing numbers of cases involving online defamation have reached the courts.

There is widespread public uncertainty about the potential risks in terms of liability for defamation in digital media contexts, but the answer from the courts appears to be fairly clear – online publication will generally be treated in a similar fashion to traditional forms of publication.

What is more contentious is the question of who, if anyone, should bear the responsibility for digital forms of defamatory publication which result not from an individual author’s activity online, but rather from algorithmic associations – most commonly in the form of a search engine result or prediction.

This has prompted calls for “algorithmic accountability” or new forms of “platform regulation” to extend upon and even challenge the place of defamation law in protecting reputation.

Professor of Law at the University of Maryland, Frank Pasquale, has argued in his book The Black Box Society that in the big data context, “authority is increasingly expressed algorithmically.”

Recently I have argued that defamation law needs to begin to engage with questions relating to data association and the big data context, and that in fact it may have a significant role to play in this area.

imageBut to harness defamation law in such a manner will require a more pragmatic and speech-protective approach from the courts and legislators.

A recent case regarding Google’s search engine has usefully brought these important questions to the attention of the High Court.

The High Court of Australia is now set to consider whether Google, as a search engine company is a publisher and whether autocomplete results are capable of being defamatory.

On June 16, 2017 the appellant Milorad Trkulja was granted special leave to appeal a December 2016 decision of the Victorian Court of Appeal, Google Inc v Trkulja.

That decision arose in the context of a successful attempt to set aside an action on the grounds that it had no real prospect of success.

In earlier cases Trkulja had been successful in claiming that he had been defamed through association with the Melbourne underworld as a result of the operation of search engines.

As the court stated in Trkulja (2016) the key questions generated by the case were “whether the search engine provider “publishes” those results for the purposes of defamation law, and whether then results produced were capable of being defamatory of the plaintiff.”

Google has in a range of such cases argued that as a matter of policy it should “be immune” from such proceedings, and though here this argument was abandoned, there is a broader policy question at stake in determining the boundaries of search engine liability.

Search engines have become the dominant (some would argue too dominant) way in which we navigate the online environment and access information.

Google dominates the market. Google resists categorisation as a media publisher, but it does appear to argue for functional protections to ensure that the free flow of information is not impeded.

In comparative contexts such as Europe these debates also play out in terms of the data protection framework and the “right to be forgotten”.

Although uncertainty remains in the Australian case law, and also in comparative contexts, this latest case makes an important contribution in several ways.

imageThe Victorian Court of Appeal emphasises that some of the difficulty in this area has been generated by comparing and drawing upon principles across a range of intermediary contexts, when in fact the law should be applied in a fact-sensitive manner, treating blogging sites, Google search engines, internet service providers, Facebook and You Tube, as different.

The court provides a thorough overview of earlier cases and concludes that a search engine is in fact a secondary publisher for whom the defence of innocent dissemination is available.

The decision also takes a sensible approach to characterising the community’s familiarity with automated processes like search engines and points to this “search literacy” and the dynamic nature of search as limiting factors in relation to questions of actionability.

From this perspective, a robust approach is taken to autocomplete predictions which are “incapable of being defamatory” and also to image association as a consequence of searching, where it is must be understood that one user’s search yields different results to another’s, and that users recognise the dynamic nature of algorithmic association in this context.

But to balance this, the court notes also that “the prospect of a publication conveying defamatory imputations in a particular instance cannot be dismissed”.

The High Court now finds itself with a significant opportunity to consider the application of defamation law principles in the context of an intermediary such as Google’s search engine.

I hope that the court will avail itself of this opportunity to clarify what has become a fairly grey and complex area in the jurisprudence.

In terms of the online context, this promises to be the most significant defamation case to reach the High Court since the seminal case of Dow Jones v Gutnick which first considered questions of jurisdiction and publication online.

As David Rolph noted of that case, the court took a fairly orthodox approach to its task in Gutnick, emphasising the bilateral quality of publication and hence the possibility of multiple publication arising in internet contexts.

According to Rolph, this resulted in evolution not revolution in terms of the development of defamation law online.

But the recent decision of the Victorian Court of Appeal, and Google’s long-running contention that it should not be regarded as a publisher, nor as a media company, point to deeper policy questions which may require a sharper approach – either in the form suggested by the Victorian Court of Appeal’s decision, or via further statutory reform of defamation law.

Earlier cases have approached the question of search engine liability in different ways.

imageFor instance, Justice Eady (pic) held in the influential English case of Metropolitan Schools v Designtechnica that Google through its search engine could not be a publisher of defamatory material generated by its algorithms.

The process was characterised in terms of its automaticity, lack of human input and as mere facilitation.

In a 2012 decision in Victoria which also related to Google and Mr Trkulja, Justice Beach found by contrast that “it was open to the jury to conclude that Google Inc was a publisher”, even without notice.

In the recent South Australian case of Duffy v Google the hurdle of automaticity was found by Justice Blue to be able to be overcome and Google therefore could be liable for algorithmic defamatory associations if notified and given reasonable time to respond.

Yet another significant strand in the mixed Australian jurisprudence is the application of the principle of proportionality by Justice McCallum in the case of Bleyer v Google, drawing on the English case of Jameel along with the rise of case management principles more broadly.

Such an approach signals a form of pragmatism that resonates with elements of the Victorian Court of Appeal’s approach in the current Trkulja case.

This also points to deeper and long standing concerns regarding the cost, complexity and chilling effects of defamation law systemically.

Chief Justice Bathurst recently remarked in an interview with Michaela Whitbourn published in The Sydney Morning Herald that there might be a need to consider further reform to defamation laws, that the costs involved are “worrying” and may not be justified by the benefits accrued in the protection of reputation.

But simplifying defamation laws is easier said than done, as evidenced by the 2005 reforms here in Australia and the vigorous and ongoing debate over defamation law reforms in the UK in 2013.

It is becoming clear that something further must be done to reform our defamation laws.

The challenge of technology, and cases involving search engines such as Trkulja (2016), point to a deeper need to reconsider what we want our defamation laws to protect, whether we need to raise the threshold in considering what might be actionable, and whether media law principles more broadly are resilient in the face of technological change.

Search engines point to broader questions of policy that will need to be addressed in the case law, but also by way of statutory reform.

*Dr Daniel Joyce is a senior lecturer in the Faculty of Law at the University of NSW, Sydney.

Photo by Niharb, supplied through Creative Commons attribution license.