Search engine liability for defamation: Duffy v Google(2)

Search engine liability

Dr Janice Duffy has been pursuing Google for redress for many years in her quest to hold Google accountable for its refusal to proactively de-index defamatory material posted on the notorious site In a recent decision, DUFFY v GOOGLE LLC [2023] SASC 13 (3 February 2023), she finally prevailed in her defamation claim against Google. In ruling for her, an Australian court confirmed that search engines can be liable for defamation when they fail to de-index snippets that are defamatory, or when snippets entice or encourage searchers to access links that are defamatory. Significantly, the court also held that a search engine that fails to implement an automatic “notice and stay down” process can also be liable for defamation.

I previously blogged about the defamation case that Dr Janice Duffy brought against Google in Australia  here and here. I also blogged about the High Court of Australia decision in Google v Defteros that also referred to the Duffy situation and which clarified the liability of search engines for defamation in Australia.

Dr Duffy’s complaint against Google

The gravamen of Dr Duffy’s defamation complaint were postings on Ripoff Reports indexed by the search engine Google that returned snippets such as the following:

Ripoff Report | Janice Duffy – Psychic Stalker! Psychics ……/Janice-Duffy…/Janice-Duffy-Psychic-Stalker-P…

The gist of Dr Duffy’s defamation complaints arose from the fact that all of the searches contained a hyperlink which, once selected, opened to the Ripoff Report webpage containing material defamatory of her.

Google did remove specific URLs on when notified by Dr Duffy, but when Google successfully removed a Ripoff Report URL containing defamatory content, it subsequently reappeared in Google search engines because Ripoff was in the habit of changing URLs specifically to avoid Google’s removal methods. Thus, her complaint was that Google was well aware of what Ripoff was doing and failed to take readily available steps to stop the wac-a-mole problem. She needed, and Google refused to provide, a “notice and stay down” process.

The snippet search results were defamatory

The court had little difficulty finding that the snippets returned by Google’s search engine had the capacity to convey defamatory meanings and in fact did so.

The ordinary and natural meaning of these snippets clearly contains the capacity to convey defamatory meanings and in fact does so, namely that Dr Duffy covertly or stealthily, sinisterly and unlawfully stalked and harassed psychics and is someone to be fearful and beware of. It may be acknowledged that the “pages of search results contain a weight of other search results” and yet the gravamen of each snippet which it must be recalled is highly ranked, is negative towards Dr Duffy. This is not a case in which refutation “destructive of the entire basis upon which the imputation relies” is present, for there is no such passage in the other snippets on the search pages containing the offending snippets.

There is however a further consideration. In Duffy v Google Inc, Blue J ascribed similar albeit not identical defamatory meanings, finding that the Reports conveyed (amongst other things) the defamatory imputations that Dr Duffy stalks and harasses psychics by persistently and obsessively pursuing them. In large measure these independently coincide with the above conclusions, although once again they are not entirely identical. The conclusions of Blue J were upheld or at least undisturbed by the Full Court in Google Inc v Duffy. In those circumstances a single judge of this Court is bound by those conclusions despite the fact that the capacity to defame was not expressly considered – albeit inconsequently – as a preliminary issue.

The court also found that the snippets were defamatory. The snippets enticed or encouraged searchers to access the defamatory content on the Ripoff pages. The court distinguished the decision of the High Court in Google v Defteros and concluded that Google was liable as a secondary publisher stating the following.

In Defteros it was authoritatively determined that a hyperlink which is not itself defamatory does not fall within the purview of publication because that hyperlink did not communicate anything, even if it links to defamatory material. This conclusion is founded on an insufficient connection between Google displaying search results in a user’s browser and the third-party website, which is hosted on that third-party URL and created by that third party. Thus, Google had merely “assisted” persons searching the Web to find and access information. It did not direct the user to a particular result and it was entirely a decision for the searcher whether to take the further step of clicking on the hyperlink so as to access the defamatory webpage…

In this instance the subject snippets did more than describe a story about a person; they described her as a furtive stalker in terms amounting to an imputation of criminal conduct, someone to be wary of and as someone to be stopped. They consisted of more than a reference or subject matter heading and the words attracted the attention and interest of the user, particularly “beware” and “stop the Australian psychic stalker”. They offered and therefore were likely to entice the user to select the accompanying hyperlink in order to clarify or to elaborate on the underlying meaning of an intriguing snippet and to seek more information. The capacity to entice the reader to select the hyperlinks was enhanced by the preeminent ranking given to the snippets by Google itself…

By taking no action until complete URLs were notified following the judgment of 27 October 2015, and by not removing the URL for RoR2 within the limitation period, Google at best acquiesced and at worst permitted access by allowing material highly defamatory of Dr Duffy to remain available on its search engines in the knowledge that it was already accessed by third parties. In other words, the only “proper inference … is that they were consenting parties to its continued presence” on the Web.

Google could not rely on the innocent dissemination defense

Innocent dissemination at common law constitutes a defence to an action in defamation. It involves the acceptance of facts sufficient to establish the legal elements of the cause of action and the avoidance of those facts by establishing a justification or excuse, as distinct from involving the denial of facts sufficient to establish an element of the cause of action.

This defence could have applied to Google’s search engine if Google could have proved that its publication was in the capacity of a subordinate distributor, that it neither knew nor ought reasonably to have known the matter was defamatory and its lack of knowledge was not due to any negligence on its part. As the court noted quoting from Google Inc v Trkulja, “a search engine is a secondary publisher of such matter, almost certainly protected by the innocent dissemination doctrine until notice is given, and for a reasonable time thereafter. Then, if it does not ‘take down’ the impugned matter, it will lose the innocent dissemination protection.”

Google was unable to make out the innocent distributor defence. It had received countless requests from Dr Duffy and her lawyers to de-index content from and knew about the prior decision of the court that held the Ripoff posts were defamatory. Yet, it while it de-indexed URLs from its search engine, it took no steps to stop their re-occurrence knowing that Ripoff would invariably just tweak the URLs to circumvent Google’s de-indexings.

It may be accepted as a matter of common sense that once having acquired the requisite knowledge, Google needed the complete URL in order to effectively block an offending website from its search engine. By the same token it cannot be accepted that with all its resources and expertise, it could not have quickly and easily ascertained the precise URL from the information it had at hand, quite apart from what Dr Duffy was providing….

The expert evidence of Mr Silver-Smith was that Google could easily deduce the complete URLs of webpages received as elided URLs, and that the screenshots of the Ripoff Reports provided by Dr Duffy contained sufficient detail to enable it to readily identify the pages in question at…

No evidence was adduced by Google suggesting that the manner of identifying complete URLs referred to by Mr Silver-Smith was impracticable or unfeasible…

Whilst ritualistically continuing to invite Dr Duffy to advise of any concerns “about the ongoing availability of any search results” and not to “hesitate to let us know”, Google thereafter rigidly maintained the stock position that “the complete URLs” were required “so that those URLs can be properly considered”. Google had long since taken the obdurate stance of insisting Dr Duffy “[i]dentify precisely … the URL of each allegedly illegal search result displayed on the Google website” before it was prepared to consider removal…

It is very apparent that Google moved to block or remove offending URLs only once given complete or unelided URL sequences. By doggedly and unrealistically insisting on nothing less despite having the means at hand to easily locate them if it had wanted to, it acted unreasonably. It was entirely reactive rather than proactive in the removals process. In truth Dr Duffy and her legal advisors (when they were engaged) were effectively stuck on a never-ending treadmill from which she could not escape of identifying complete URLs, securing removal by Google only to find the same posts with altered URLs inexorably reappearing whilst Google stood by doing nothing itself. It is to be recalled in this context that the evidence of Mr Silver-Smith was that “Google could have fairly easily created a method to detect Ripoff Report’s subversion, and automatically counteracted it”, evidence which Google at no point countered or disproved….

In the above circumstances Google singularly fails to prove that it did not or that it could not reasonably have known the matter was defamatory and that its lack of knowledge was not due to any negligence on its part.

For the reasons advanced above it must follow that the defence of innocent dissemination is unavailable to Google.

Concluding remarks on search engine liability for defamation

The Duffy case once again confirms that, at least in Australia, search engines can be liable for defamation as secondary publishers. Search engines can be liable when they fail to de-index snippets that are defamatory, or when snippets entice or encourage searchers to access links that are defamatory. Search engines are protected by the innocent dissemination doctrine until notice is given, and for a reasonable time thereafter. Then, if they do not disable access to the impugned matter, they will lose the innocent dissemination protection.

What is significant about the Duffy decision is the court’s holding that once a search engine has a certain degree of knowledge that the site publishing the defamatory material is subverting prior de-indexed URLs, the search engine can also be liable for defamation unless it takes measures to automatically counteracted it. In effect, the court endorsed a “notice and stay down” requirement for search engines seeking immunity for defamation liability.

What’s next for Dr Duffy?

The litigation against Google is not over. Google may once again appeal the unfavorable decision against it. Dr Duffy represented herself at trial against Google and may be required to continue the fight herself in an appeal to the Australia Full Court (the Court of Appeal). There is also the matter of the remedy the court will award her including whether the court will grant an injunction against Google, the scope of the injunction, what damages she will recover and whether a large sum will be awarded as additional or punitive damages,

For more on the case, see the blog post by Neil Turkewitz.

** Image, Source – SEL.

Print Friendly, PDF & Email

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: