Introduction
The High Court has paved the way for an individual to sue Google for allegedly defaming him in a series of search results and auto-complete phrases.
Background
Milorad Trkulja (aka Michael Trkulja) commenced proceedings against Google Inc (later substituted with Google LLC) (Google) in which he alleged that Google defamed him by publishing images which conveyed certain imputations, including that Mr Trkulja was a “hardened criminal”. Mr Trkulja claimed that Google published images and online content to people who used Google’s image search function and its web search function to search for “Milorad Trkulja” or his alias “Michael Trkulja” during the period 1 December 2012 and 3 March 2014, and that those publications gave rise to the alleged defamatory imputations. The material included, for example, Google’s “autocomplete” function which predicted phrases such as “michael trkulja melbourne underworld crime”; and search results which returned images of Mr Trkulja amongst images of convicted Melbourne criminals.
Google applied to the Supreme Court of Victoria to have the statement of claim set aside on the basis that:
- Google is not a publisher and did not publish the images or online content;
- the particular search results underpinning Mr Trkulja’s claim were not defamatory; and
- Google was entitled to immunity from suit.
The primary judge dismissed Google’s application on the grounds that there was no proper basis for concluding that Mr Trkulja had no real prospect of establishing that any of the material he complained of was defamatory.1
On appeal, the Victorian Court of Appeal overturned the primary judge’s decision and concluded that the material was incapable of conveying any of the alleged defamatory imputations.
The decision of the High Court
The High Court unanimously found that the Court of Appeal had fallen into error in several respects.
First, the Court of Appeal had applied the wrong test for capacity to defame; it resolved that the question for determination was whether “any of the defamatory imputations which [were] pleaded [were] arguably conveyed”2. Given that the proceeding was in the summary jurisdiction, the correct test was “whether any of the search results complained of [were] capable of conveying any of the defamatory imputations alleged.” 3
In asking the wrong question, the Court of Appeal led itself into error and, despite the summary nature of the application, made mixed findings of fact and law. Notably, the Court of Appeal concluded that:
- a search engine such as Google was a publisher of search results (including the autocomplete predictions);
- the defence of innocent dissemination will almost always be available before the publisher is notified of an alleged defamation; and
- a user who searched phases such as “michael trkulja melbourne underworld crime” would “very probably perceive a disconnect between the images and the search terms” and “would recognise … that the search results in their entirety did not reflect the meaning of the inputted words considered as a phrase”.4
The Court of Appeal also considered that the results for an image search and the results for a web search were single composite publications, even though Mr Trkulja had pleaded that each search and corresponding result was a separate publication since a person conducting one of the searches may not have conducted any of the other searches. The High Court held that Mr Trkulja’s approach was consistent with previous authority and that, if the matter were to proceed to trial, the difference could prove significant.
Without the benefit of discovery, opposing evidence and cross-examination, the Court of Appeal was not in a position to make substantive findings as to Google’s capacity to be a publisher, the availability of possible defences or the standard of knowledge of a reasonable user of a search engine.
Moreover, the High Court agreed with the primary judge that it was “evident” that at least some of the material had the capacity to convey one or more of the defamatory imputations. In arriving at that conclusion the High Court noted that it could be assumed that the fact that a person who searched for members of the Melbourne criminal underworld would rationally think that there was a connection between that search and the corresponding search results. Further, unlike some of the persons whose image appeared in the search results (such as an image of Marlon Brando in his role as “the Godfather”), Mr Trkulja was a relatively unknown person and so it would be open to the person conducting the search to conclude that Mr Trkulja was connected with the criminal underworld.
Wider implications of the decision
Although the High Court did not make any definitive findings as to Google’s capacity to be a publisher of defamatory matter, there are some significant principles that emanate from the judgment.
First, the decision establishes that search results and auto-predict functions of a search engine can be capable of conveying defamatory imputations.
Second, there is a strong basis for concluding that, where a search engine intentionally participates in the communication of allegedly defamatory search results, it can be a publisher of defamatory matter. On this point, the High Court emphasised that it is not necessary for an applicant to plead that the search engine is a primary or secondary publisher, given that all degrees of publication are sufficient to ground liability as a ‘publisher’.
Third, the question of whether the words or images complained of are capable of conveying a defamatory imputation is a question of law about which reasonable minds may sometimes differ and, accordingly, a finding that they are incapable of bearing a defamatory imputation should only be made with “great caution”.
The High Court observed that the standard of knowledge and comprehension of a “reasonable” user of a search engine has yet to be authoritatively determined, but alluded to the possibility that there may be more than one “class” of users for the purposes of defamation; for example, first-time users of Google and experienced participants.
The High Court also emphasised that, in relation to search engines/results, alleged imputations must be considered not only in the context of search results, but also against the background of search terms. It also endorsed considering each search and corresponding search result together, but separate from any other search and corresponding search result.
Lastly, their Honours affirmed the test for whether a published matter is capable of being defamatory as being what ordinary reasonable people would understand by the matter complained of and, in making that assessment, it is necessary to bear in mind that ordinary men and women have different temperaments and outlooks, degrees of education and life experience.
Mr Trkulja’s substantive defamation proceeding has yet to go to trial, but other search engines and competitors of Google will undoubtedly be interested to see what effects it might have on their liability in publishing potentially defamatory content through search results.
1.Trkulja v Google Inc [2015] VSC 635. (In an earlier, related, proceeding based on similar facts, Google had contended that it could not be liable to Mr Trkulja for its wholly automated publications (i.e. search results): Trkulja v Google Inc LLC & Anor (No 5) [2012] VSC 533. See the discussion in our Focus Paper of 20 August 2014 “Freedom of search results: Google and others not liable for defamatory search results, at least until they are put on notice”).
2. Trkulja v Google LLC [2018] HCA 25 at [42], discussing Google Inc v Trkulja (2016) 342 ALR 504 at 597; [2016] VSCA 333 at [388]-[390].
3. Trkulja v Google LLC [2018] HCA 25 at [52].
4. Google Inc v Trkulja (2016) 342 ALR 504 at 537; [2016] VSCA 333 at [151].
5. Dow Jones & Co Inc v Gutnick (2002) 210 CLR 575 at 652-653; [2002] HCA 56 at [197]-[199]. See also The Buddhist Society of Western Australia Inc v Bristile Ltd [2000] WASCA 210 at [10].
Liability limited by a scheme approved under Professional Standards Legislation.
© ADDISONS. No part of this document may in any form or by any means be reproduced, stored in a retrieval system or transmitted without prior written consent. This document is for general information only and cannot be relied upon as legal advice.