The Supreme Court of New South Wales has significantly limited the liability of search engines, such as Google, for any defamatory ‘publication’ which arises from an automatically generated search result on the site. In doing so, the Court has declined to follow an earlier Victorian decision and instead favoured a series of English cases concerning defamatory search results.
In Bleyer v Google Inc  NSWSC 897, the Court found that Google had not ‘published’ search results created by its wholly automated processes, and before any take down complaint had been received. Judgment in the case was delivered last week by McCallum J who also confirmed the ability of the Court to stay or dismiss, as an abuse of process, an action on the grounds that the resources to be expended to determine the claim would be out of all proportion to the interest at stake. In this case, Mr Bleyer’s proceedings were permanently stayed.
Mr Ronald Bleyer had brought an action against the US based Google Inc regarding Google search results for his name. He alleged that seven of Google’s “snippets” (the short excerpts of web pages displayed in search results) either solely or in combination with the full web pages identified and to which they were hyperlinked, defamed him. He could particularise publication to only three people in Australia.
The Court accepted that Google has no editorial input into its search results, apart from designing pre-programmed algorithms that crawl the web, automatically generating snippets and hyperlinks in response to search queries. Nor does Google have control over the content of third-party webpages.
Google does, however, operate a ‘notice and take down procedure’ whereby it removes, from its results, URLs that are the subject of complaint and in violation of applicable laws. Mr Bleyer’s solicitor had put Google on notice after the results had been seen by two of the people mentioned in particulars. However, before the process had been completed and some of the offending material removed, publication to the third person had taken place.
Trkulja, and publication by automation?
On this basis Google argued that, having no human input into the search process, it could not be regarded as having ‘published’ its search results for the purposes of defamation law. It relied on a line of English cases, decided by Eady J,1 for the proposition that persons, including search engine providers in respect of search results “…who truly fulfil no more than the role of a passive medium for communication cannot be characterised as publishers…”.2 On that approach, snippets were not normally considered to have been authorised by Google in any meaningful sense, they were merely part of Google’s facilitative role via the provision of its search service. Accordingly, Google could not be a ‘publisher’ at common law.3
As discussed below, the situation was acknowledged to be different in the event that Google was put on notice as to possibly defamatory material.
Those English decisions had previously been considered and rejected by the Supreme Court of Victoria in 2012, in the Trkulja case.4 In those proceedings, Mr Trkulja claimed that Google had published two classes of material defamatory of him through search results, being: image results; and web results. A jury found that Mr Trkulja was entitled to damages in respect of the images. This was on the basis that Google had made those results available over a certain time period when it was open to the jury to find that Google had not established a defence of innocent dissemination.5 Google appealed on the basis that it could not be a ‘publisher’ of either the images or the web matters.6
Both parties accepted that it was necessary to establish that Google intended to publish the relevant material.7 Responding to Google’s non-publication argument based on its search engine being fully automated, Beach J held that:8
“ … it was open to the jury to … conclude that Google Inc was a publisher even before it had any notice from anybody acting on behalf of the plaintiff. The jury were entitled to conclude that Google Inc intended to publish the material that its automated systems produced, because that was what they were designed to do upon a search request being typed into one of Google Inc’s search products. In that sense, Google Inc is like the newsagent that sells a newspaper containing a defamatory article.”
No human input, no notice? No publication.
In contrast to Trkulja, McCallum J took the view that previous Australian authority – concerning a municipal authority’s failure to remove defamatory posters placed in bus shelters – did not support a finding that Google was liable for publication before being put on notice by the plaintiff.9
In company with the English Court of Appeal in Tamiz (decided after Trkulja), her Honour also rejected the idea that Google has the same role as a secondary publisher or distributor in providing search results. Rather, persons involved only in dissemination are not to be treated as publishers unless they know that the publication is likely to be defamatory.10
Given her characterisation of Google’s search engine as receiving no human input apart from the initial algorithm creation, her Honour was unable to accept that this alone could ground liability as a ‘publisher’ at common law.11 Accordingly, “at least prior to notification of a complaint … Google Inc cannot be liable as a publisher of the results produced by its search engine”.12 With there being no publication in this period, the ‘defence’ of innocent dissemination could not arise.13
Staying or dismissing proceedings as an abuse of process: disproportionality.
In Mr Bleyer’s case, there remained the question of Google’s liability for the publication to a single person of defamatory search results after it was put on notice of their content.
Again on the basis of English authority, Google submitted that the proceedings should be stayed or summarily dismissed, given the disproportionality between the resources required to determine the claim, and the vindication sought. This possibility was recognised in England in Jameel (Yousef) v Dow Jones & Co Inc  EWCA Civ 75. Mr Bleyer contended that Jameel should not be followed on the basis that it was overly reliant on the UK’s Human Rights Act 1998 (and thus the European Charter of Human Rights). Moreover, the plaintiff argued it should not be applied so early in the proceedings, before the benefit of further and better particulars after discovery and interrogatories.
McCallum J rejected the latter argument on the basis that the processes of the Court should not be beholden to speculation, especially where so few downloads had been identified.14 Although the plaintiff could have identified further facts from which to infer a wider publication, it had not done so: there was no basis, beyond speculation, that more than three people had ever accessed the search results.
McCallum J accepted that it was not necessary to refer to English law to determine what an abuse of process was in the New South Wales’ Supreme Court. However, her Honour was satisfied that sufficient disproportionality of the type contended could – in rare occasions – amount to such an abuse, and lead to a stay or dismissal of an action.15
On these facts, with only one possible publication to one person, in circumstances where a judgment is likely to be unenforceable against the publisher in the United States, and estimating the resources that would be expended by the Court and parties; the proceedings were ordered to be permanently stayed.
- It remains arguable that search engine providers may be liable for defamation through search results after notification. So, timely invocation of notice and takedown procedures can be crucial.
- Conversely, search providers and others providing information through automated processes can breathe more easily, but ought to be assiduous in following up complaints procedures.
- It will become increasingly difficult to justify speculative proceedings for a plaintiff based only on a handful of downloads in the hope that more evidence will later come to light.