DR YEUNG, SAU SHING ALBERT v. GOOGLE INC

HCA 1383/2012

IN THE HIGH COURT OF THE

HONG KONG SPECIAL ADMINISTRATIVE REGION

COURT OF FIRST INSTANCE

HIGH COURT ACTION NO 1383 OF 2012

____________

__________________

D E C I S I O N

__________________

I. INTRODUCTION

1. On 5 August 2014, I handed down my decision in this action in respect of a summons under Order 12 rule 8 of the Rules of the High Court (“RHC”) and/or the inherent jurisdiction of the court, which was heard before me on 8 May and 7 November 2013 (“Decision”).  For the sake of convenience, I shall adopt the abbreviations in the Decision.

2. By the Summons, Google Inc sought inter alia the following reliefs:

(a) a declaration that the court has no jurisdiction over Google Inc in respect of the subject matter of the claim and/or the reliefs/remedies sought against them in the present action, or alternatively a declaration that the court should not exercise any jurisdiction it may have;

(b) an order that (i) the CWoS and Leave Order be set aside or discharged, (ii) service of the CWoS on Google Inc be set aside, and (iii) all subsequent proceedings in the present action be stayed.

3. By the Decision, I dismissed the Summons and granted a costs order nisi in favour of Yeung against Google Inc (with all costs reserved if any) to be taxed if not agreed (“Order”).  There was no application by either party to vary the above costs order nisi, so it has been made absolute.

4. On 12 September 2014, the plaintiff filed a summons for leave to appeal against the Order (“Leave Summons”) with a draft Notice of Appeal (“Draft Notice”) annexed thereto.  The hearing of the Leave Summons came before me on 23 October 2014 (“Leave Hearing”).

II.  LEGAL PRINCIPLES

5. The Order is an interlocutory order, and it is incumbent upon Google Inc to persuade this court to grant leave to appeal.  Under section 14AA(4) of the High Court Ordinance Cap 4, leave to appeal shall not be granted unless the court is satisfied that the appeal has a reasonable prospect of success or there is some other reason in the interests of justice why the appeal should be heard.  In SMSE v KL,[1] Le Pichon JA pointed out that such leave to appeal is not lightly granted.  Reasonable prospect of success involves the notion that the prospect of succeeding must be “reasonable” and therefore more than “fanciful”, without having to be “probable”.[2] Furthermore, it is pertinent to bear in mind that even if there is a reasonable prospect of success on appeal, the court still retains a discretion whether to grant leave to appeal, although the fact that there is, ex hypothesi, a reasonable prospect of success would heavily influence the court’s exercise of discretion.[3]

III.  DRAFT GROUNDS OF APPEAL

6. In the Draft Notice, Google Inc relies on four draft grounds of appeal.  For the present purpose, Mr McCoy SC (and Ms Ngai with him), counsel for Google Inc, in his written skeleton submissions condensed them into three broad grounds, which I summarise as follows:

(a)  Ground 1 – no publication to third parties: Google Inc claimed I have erred in holding there is a good arguable case that viewing, downloading, printing the offending Words from the Websites by users at the IT Dept and ILL constitute “publication” to a third party (“Ground 1”);

(b) Ground 2 – Google Inc was not the “publisher”: Google Inc claimed I have erred in holding there is a good arguable case that they are a “publisher” on the basis that Google Search may not be just a passive facilitator of information vis-à-vis the Autocomplete and Related Search features (“Ground 2”);

(c)  Ground 3 – failure to consider context: Google Inc claimed I have erred in failing to consider the neutral context in which the Words appeared, and in holding there is a good arguable case of defamation against Google Inc to seize it of jurisdiction under Order 11 rule 1(1)(f) and rule 14 of the RHC (“Ground 3”).

7. In respect of Ground 1, Mr McCoy SC submitted that The Duke of Brunswisk v Harmer[4] was outmoded and Jameel (Yousef) v Dow Jones & Co Inc[5] relevantly applied, so publication of the alleged defamatory material to Yeung’s servants and agents tasked to search/print the Words (ie the users at the IT Dept and ILL) did not demonstrate genuine publication of the alleged defamatory Words in the libel sense to third parties, and a stay of proceedings was therefore appropriate.  In advancing such argument for the purpose of the Leave Summons, Mr McCoy SC relied on the judgment of McCallum J of the Supreme Court of New South Wales in Bleyer v Google Inc[6] which I will return to below.  Mr McCoy SC suggested there was a lack of sufficient or material publication to a third party to support service of process on Google Inc out of jurisdiction, and I should have concluded it would be a disproportionate engagement of the court’s resources to allow Yeung’s artificially constructed “set piece” alleged libel to proceed.

8. In respect of Ground 2, Mr McCoy submitted my characterisation of Google Search in the Decision was erroneous.  He argued that since I had accepted Google Inc’s algorithms weighed several factors outside their control and Google Inc had no hand in what users had searched in the past or the content of web information, it was wrong for me to conclude there was an arguable case that Google Inc had “actively facilitated or intentionally assisted” in the process of conveying the alleged defamatory Words.  Mr McCoy SC submitted that Google Search’s recombination and aggregation of data was merely for the purpose of improving the search function for its users, and did not affect the underlying information dictated by (amongst other factors) search queries and web content made by other users.  Mr McCoy SC complained that in holding there was a good arguable case that Google Inc was a “publisher”, I essentially made them a victim of their own success in improving the quality of the information generated by Google Search.  He added that Google Inc’s ability to censor material by manually editing results was neither here nor there to the issue of whether the Autocomplete and Related Searches functions fell within the passivity rule.

9. Mr McCoy SC relied on McCallum J’s observations in Bleyer that led her to reject Trkulja v Google Inc LLC & anor (No 5),[7] which he said cast “immense doubt” as to whether Trkulja v Google Inc LLC was correct in law, and I will return to this below.  Mr McCoy SC contended McCallum J’s observations demostrated a good arguable case that Google Inc could rely on the passivity rule for their Autocomplete and Related Searches functions.

10. In respect of Ground 3, Mr McCoy SC submitted that upon proper understanding of the functionalities of Google Search, it was plain the Words were merely a true representation of user activity without Google Inc having judged the correctness, credibility or reliability of the contents (which contention harked back to Ground 2).  Since the neutral context in which the Words appeared would be paramount, Mr McCoy SC suggested I should have held there was no arguable case against Google Inc for defamation to seize it of jurisdiction under Order 11 of the RHC.

IV.  GROUND 1

11. Mr McCoy SC is correct to say that in the Decision I have accepted the principle in The Duke of Brunswick and Pullman v Walter Hill & Co Ltd[8] that publication to at least one other person satisfies the cause of action for defamation, but in the Decision I have gone further to consider and discuss Mr McCoy SC’s core complaint of lack of publication beyond users at the IT Dept and ILL and the impact of such complaint on service out of jurisdiction before concluding Yeung has a good arguable case on the issue of “publication”.

12. Mr McCoy SC has suggested that the facts in Bleyer are almost identical to the present case, and has reminded that the outcome in that case is the opposite result.  In that case, Mr Bleyer alleged that Google Inc published seven defamatory items (mainly “snippets” and hyperlinks to offending articles) about him to three people who performed “google” searches.  Two readers accessed the materials in December 2012.  Mr Bleyer’s solicitors asked Google Inc to remove the materials in March 2013.  Another reader accessed the materials in April 2013.  Google Inc took down some but not all of the materials in May 2013.  Google Inc applied to stay the proceedings or dismiss them as an abuse of process.

13. Mr McCoy SC has drawn my attention to the following observations by McCallum J in Bleyer:

“29.  In the present case, according to the existing pleading, the “best particulars of publication” the plaintiff can presently provide is that the seven matters complained of were published to two people who live in the state of Victoria (supplemented in correspondence by a third, a resident of this State).  In my view, where the prospect of further and better particulars being provided after discovery and interrogatories is necessarily speculative, the appropriate course is to determine the defendant's application on the pleadings and particulars as they stand.  The processes of the Court should not be beholden to speculation.  In defamation proceedings in particular, the scope of publication is an important issue.  The Court should not encourage a practice of commencing proceedings for defamation on the strength of a small number of downloads of material from the Internet in the speculative hope of proving wider circulation after discovery and interrogatories.

……

31. One of the difficulties with a defamation claim based on the publication of search results using Google or other search engines is that the relevant audience is necessarily confined to the class of persons who have undertaken a Google search using terms that identify the relevant defamatory web page according to the search engine's algorithms. There is no basis in the pleading or in the evidence on the present application (beyond speculation) for apprehending any likelihood that there are more than three people in that class in the present case.”

14. But Mr Ng, counsel for Yeung, has reminded that in between paragraphs 29 and 31 of the judgment in Bleyer, McCallum J also said as follows:

“30. That is not to say that the number of actual downloads identified in any individual case will always be determinative as to the likely scope of publication. It is always open to a plaintiff to identify additional facts, matters and circumstances on the strength which a wider scope of publication may be inferred, such as the popularity of the relevant website, the public profile of the plaintiff and any other factor pointing to the likelihood of the defamatory web page being downloaded by a broader range of people. No such particulars are given here.” (my emphasis)

Mr Ng has argued that the present circumstances are a far cry from the situation in Bleyer where the plaintiff did not provide any particulars from which a broader range of publication could be inferred. 

15. In my view, Bleyer plainly acknowledges that publication can be inferred if appropriate factual foundation for such inference to be drawn is laid before the court.  In that case, the plaintiff failed to identify further facts from which to infer a wider publication, so there was no basis, beyond speculation, that more than three people had ever accessed the search results.  But here, in Part VIII(a) and (c) of the Decision, I have discussed the legal principles drawn from relevant authorities as well as the contextual background of the alleged defamatory Words, including the public profile of Yeung, his involvement in the entertainment business, the subject matter of the Words possibly being of continuing interest, the popularity/commonplace nature of “googling” about persons, and the ready accessibility of the Autocomplete and Related Searches features to internet users (which are the very factors McCallum J referred to in Bleyer for drawing inference as to the likely scope of publication), before concluding it is an open question of fact, by inference or otherwise, that there may have been real and substantial publication. 

16. Indeed, in the Decision I have gone further to analyse the inference that arguably can be drawn from the statistical information in the Yuen Aff which has led me to the view that had there been material non-disclosure at the ex parte stage (which I do not agree) I would not have set aside the Leave Order or alternatively I would have re-granted the same. 

17. Mr McCoy SC has not persuaded me the present context is, as he has suggested, identifical to that of Bleyer where the plaintiff simply produced no evidence upon which an inference could be made of substantial publication of the offending words.

18. Next, Mr McCoy SC in his written submissions for the Leave Hearing has emphasised it will be a disproportionate engagement of the court’s resources to allow Yeung to proceed with his libel claim based on mere publication to users of the IT Dept and ILL. 

19. I note Bleyer was ultimately decided on the issue of proportionality, ie whether the case is worth running given the potential costs and outcome, or to put it in another way, whether the game is worth the candle.[9] McCallum J accepted it was not necessary to refer to English law to determine whether there was an abuse of process, and she was satisfied there was sufficient disproportionality of the type contended (ie between the resources required to determine a claim and the interest at stake) that could on rare occasions amount to an abuse of process, and thereby lead to a stay or dismissal of the action.  However, the learned judge accepted the value of interest could include considerations other than monetary value, eg vindication of reputation. 

20. But McCallum J’s conclusion in Bleyer that the legal costs and court resources required to determine the claim would be “out of all proportion” to the interest at stake was necessarily informed by the absence of any inference of wider publication in that case, which is to be contrasted with the discussions in the Decision on inference as to publication in the present context.  In any event, in the Decision I have also considered the balance of interests between the scale of publication and the vindication of reputation, and between freedom of speech, press and publication and respect for rights/reputations of others before coming to the view that Yeung has put forward a good arguable case.  In the circumstances, I do not agree with Mr McCoy SC that the conclusion in Bleyer based on its own circumstances obviously changes the scene. 

21. But that said, I bear in mind the novelty of the arguments on Jameel abuse of process and on staying/dismissing proceedings on the proportionality principle in this jurisdiction.  It is true that locally there has been some previous judicial reluctance to follow Jameel (Yousef), but there is no judicial guidance at all on how the proportionality principle features in the post-CJR landscape.  Indeed, according to Bleyer, there has been debate even in the Australian jurisdiction as to whether Jameel abuse of process is applicable, and McCallum J’s approach premised on the non-Jameel proportionality principle has not been adopted in Australian courts before.  In such context, I see the force of Mr McCoy SC’s submissions that oversight (and more significantly, guidance) by the Court of Appeal on the balancing of the factors, especially in the post‑CJR regime,  that may encourage courts to put the brakes on costly and time-consuming defamation actions will be relevant and useful.  I am persuaded this comes within the “some other reason in the interests of justice” limb for granting leave to appeal.

V.  GROUND 2

22. Mr McCoy SC has laid emphasis on what I have described in the Decision as the passivity rule on the basis of the functionalities of Google Search which, according to Google Inc, demonstrate that it is a neutral tool.  McCallum J in Bleyer accepted that Google Inc was not a publisher of the alleged injurious materials for the purpose of defamation law.  Her Honour declined to follow Trkulja v Google Inc LLC and chose to follow the English authorities on the basis there was no human input in Google Search apart from setting it up in the first place, and “at least prior to notification of a complaint …… Google Inc cannot be liable as a publisher of the results produced by its search engine”,[10] and hence the question of innocent dissemination did not arise. 

23. In light of the different approaches adopted by Beach J in Trkulja v Google Inc LLC and McCallum J in Bleyer, which are both first instance decisions, it is perhaps correct for Mansfield J to say in Rana v Google Australia[11] that whether or not an internet search engine can be considered a publisher in defamation law “is not settled in Australia”.  I am not persuaded McCallum J has said the last word in this evolving area of the law at least as applicable to this jurisdiction, and Trkulja v Google Inc LLC and A v Google New Zealand Ltd cannot be so lightly brushed aside.

24. Moreover, McCallum J followed Bunt v Tilley & ors,[12] Metropolitan International Schools Ltd v Designtechnicia Corporation[13] and Tamiz v Google Inc,[14] but in this jurisdiction the Court of Final Appeal in Oriental Press Group Ltd & anor v Fevaworks Solutions Ltd[15] considered the Bryne v Deane approach unsuitable for dealing with internet intermediaries because they do participate in a real sense in the dissemination of the defamatory material (even though the defence of innocent dissemination is available to secondary publishers conditional on them proving that they did not know and did not have reasonable grounds to be aware of the defamatory matter).  Fevaworks Solutions Ltd has expressly parted ways with Tamiz on a matter of principle; our Court of Final Appeal has rejected the uncomfortable conflation of the innocent dissemination defence applicable to the period before Google Inc became aware of the defamatory material and the Bryne v Deane approach applicable to the period after notice was received.  Mr McCoy SC in his oral submissions at the Leave Hearing has confirmed he is not saying that Fevaworks Solutions Ltd was wrongly decided.  That being the case, question immediately arises as to the utility (or the extent of the utility) of Bleyer in this jurisdiction on the matter of whether Google Inc is a “publisher” when viewed through the prism of the fundamental principles in Fevaworks Solutions Ltd.

25. Mr McCoy SC has suggested that even though Fevaworks Solutions Ltd has adopted the strict publication rule for Hong Kong, the Court of Final Appeal could hardly have envisaged its applicability to the significantly different and unusual context of the present action.  Mr McCoy SC has reminded that an internet search engine (with emphasis on lack of human input for its operations) is sharply different from an ISP that hosts an internet blog/forum (eg Fevaworks Solutions Ltd) or that creates an internet hyperlink (eg Crookes v Newton[16]), and this must be an energising reason that articulates the decision in the present case.

26. But one cannot forget the alleged defamatory materials in Bleyer were mainly “snippets” and hyperlinks.  Whilst it may be argued that the latter functions merely as neutral location tools that indicate where existing internet content can be found (which arguably makes the analogy to mere conduit or passive facilitator more compelling),[17] the fundamental principles in Fevaworks Solutions Ltd (especially in light of their divergence from the English authorities) make the analogy less persuasive in respect of “snippets” since an ISP does deal with content when generating the “snippets” that point to the underlying internet materials by reference to their content such that, quite irrespective whether or not the underlying web materials are defamatory, it is arguable the publisher of any defamatory content of “snippets” can only be the ISP.

27. Here, there cannot be any dispute that Google Search by its Autocomplete and Related Searches features do not merely feed all historical user search queries and/or all historical web content to the internet user who makes a “google” search.  Instead, Google Search’s algorithms (including PageRank) designed by Google Inc decide ranking by the number/quality of the incoming links from other pages, and measure the popularity, demand and/or usage of the historical search terms, and ultimately aggregate and reconstitute such content and terms into a few keywords that appear as Autocomplete and Related Searches predictive suggestions.  So whilst the algorithmic processes are automated, Google Inc deploys artificial intelligence to amass information from search queries and web content, and then present them to the users as Autocomplete and Related Searches predictive keywords.  It is on such basis that I have found it may arguably (but not necessarily or definitively) be said Google Search does deal with content by crawling and indexing websites, aggregating data from previous search queries, and generating predictive keywords for the Autocomplete and Related Searches features, and by virtue of such features it is also arguable that unlike mere conduits or passive facilitators Google Inc may be considered to be a publisher of injurious content of the predictive suggestions.[18]  I also bear in mind that it is on general and broad principles that Fevaworks Solutions Ltd has held that secondary publishers who do not actually know and do not intend to publish the defamatory materials are nevertheless “publishers” under the strict publication rule,[19] but they can avoid liability by invoking the innocent dissemination defence (if they can).

28. Mr McCoy SC has accepted that Bleyer may not be definitive in this jurisdiction.  But that said, he may have a point in his submissions that nuanced principles are likely to develop in a “different context” in the internet world (in relation to Autocomplete and Related Searches features) within the framework identified by Fevaworks Solutions Ltd.  I do not ignore (a) the possibility that Hong Kong courts may have to explore and discover our own way instead of transposing foreign case law on the issue of whether Google Inc is a “publisher” in the present context, and (b) the risk posed by the interplay of divergent views in the common law jurisdictions and the binding broad principles in Fevaworks Solutions Ltd that may make it easy for a first instance court to traipse down a wrong path.

29. Mr Ng has suggested that I should let the present action proceed and take its course for the factual circumstances to be elicited at trial in order to properly inform any development in internet defamation law in such “different context”, and he has further reminded that after all Yeung is only required to show a good arguable case that Google Inc was a publisher of the Words.  But having considered the arguments by Mr McCoy SC and Mr Ng, and the matters discussed in Part VII below, I have come to the view that leave to appeal should be granted on the “some other reason in the interests of justice” limb so that the Court of Appeal can impart their wisdom on the implications of divergent views in international jurisprudence for this jurisdiction.

30. Finally, Mr McCoy SC has suggested that Google Inc’s ability to censor materials by manually editing results is neither here nor there.  I am not particularly persuaded by such argument.  Indeed, recent demonstration of willingness on the part of Google Inc to undertake the massive task of dealing with removal requests to comply with the decision of the Grand Chamber of the European Court of Justice involving Google Spain SL[20] arguably supports the conclusion that “there is room for factual debate as to what further blocking steps, if any, would be open for Google Inc to take or how effective they may be”, which also adds colour to the argument that Google Inc may be viewed as a “publisher” of the Words.

VI.  GROUND 3

31. I am not persuaded there is merit to Ground 3 for Google Inc has applied under the Summons on the basis that they do not take issue that the Words are defamatory of Yeung.

VII.  SOME OTHER REASON IN THE INTERESTS OF JUSTICE

32. I have touched on this matter in the above discussions.  But before I say anything more, I shall briefly deal with one short matter.  Mr McCoy SC has urged this court to avoid a mechanistic approach, and not to assume that after the CJR first instance courts have to adopt an austere approach in exercising their gatekeeper function on leave to appeal. 

33. I am quite unconvinced that after the implementation of the CJR judges have been parsimonious in granting leave to appeal.  The busy diary of the Court of Appeal proves otherwise.  I pause here to emphasise the exercise of the jurisdiction to give or to refuse leave to appeal by the court below, which serves a filtering purpose, should not be overly generous; otherwise unmeritorious challenges would add to costs to be incurred by the opposing parties and would waste court resources.  Lam VP in Singh Arhun by his next friend Singh Anita Guruprit v The Secretary of Justice on behalf of The Commissioner of Police & anor[21] recently reminded litigants of the function of the first tier application for leave to appeal as follows:

“10. …… a first tier application before the matter is brought before the Court of Appeal serves an important function. Not only is this function valuable to this court, it should also be of value to the parties. The party applying for leave should obviously reflect on the arguability of his grounds if the lower court refuses leave. At the same time, the opposing party should also review the merit of the application in light of the contentions run at the first tier hearing. This should assist the parties in formulating and responding at the second tier stage if an application is made to the Court of Appeal for leave. In other words, this court expects the submissions advanced in the second tier level to be more concise and focused.”

34. Turning back to the arguments raised by counsel, Mr McCoy SC has argued that it cannot be disputed the issue before the court is a novel, developing one which involves a rapidly evolving area of the law.  He has further reminded that I have described this as a “troubling grey area” in the Decision.  He has submitted that the appeal should be heard in the interests of justice because important points of law arise in the context of developing and inconsistent decisions that have recently mushroomed in the common law world, and despite such jurisprudence the Decision is nevertheless a foray into virgin legal territory without any appellate guidance in local and other common law jurisdictions as to whether Autocomplete and Related Searches features are amenable to the tort of defamation. 

35. Mr McCoy SC has reminded that the question of service out of jurisdiction goes to the fundamental matter of the court’s own jurisdiction, and appeal cases on such question abound in the law reports both in this jurisdiction and abroad.  He says the present case also touches on the question as to whether Yeung’s cause of action under the tort of defamation is justiciable in the search for “the comfortable equilibrium”[22] between the limits of the tort of defamation vis-à-vis an internet search engine and the competing interest of freedom of expression and publication.  Mr McCoy SC says that against such backdrop this court should instantly grant leave to appeal, and he fears that without approval/disapproval of the Decision by the Court of Appeal it will become lost in what he has described as the “wilderness of the first instance judgments”.

36. On the other hand, Mr Ng has submitted the fact that this is a rapidly developing area of the law does not aid Google Inc because the criteria for obtaining leave to serve out of jurisdiction merely requires a plaintiff to show he has a good arguable case that falls within the ground(s) under Order 11 rule 1(1) of the RHC, and to satisfy the court there is a serious question to be tried on the merits of the claim.  Indeed, Mr Ng has submitted that rapid legal development provides all the more reason for any novel points to be determined in their factual context to be elicited at trial.  He reminds that Fevaworks Solutions Ltd concerned an appeal after trial rather than an interlocutory ruling, and the appellate courts in that case had the benefit of the factual findings made by the first instance court to put their exposition of the law in context.

37. Even though I can see the force of Mr Ng’s arguments, and (as seen from the above discussion) I am not quite persuaded that Bleyer necessarily raises reasonable prospects of tilting the balance against my conclusion that Yeung has established a good arguable case, I have come to the view that it is inappropriate for this court to make an unseemly spectacle of standing in the way of an appeal that hopefully will bring enlightenment by the Court of Appeal in this new and uneasy area of defamation law.

VIII.  CONCLUSION

38. In the premises, I consider leave to appeal ought to be given.  I grant an order in terms of paragraphs 1 and 2 of the Leave Summons.  Both counsel have agreed that should leave be granted, costs should be in the cause of the appeal, hence the costs order made is an absolute and not nisi one.