Court of Appeal decision in Gilroy & Byrne v O’Leary and Google [2025] IECA 42
27 February 2025
Introduction:
The Statute of Limitations Act 1957 provides that an action alleging defamation must be commenced within 12 months of the defamatory material first being published. This period may be extended for a further 12 months at the discretion of the Court (but in reality, it rarely is).
In simple terms, if a newspaper publishes a defamatory statement, the cause of action accrues on the day that newspaper goes on sale, and proceedings must issue within 12 months. In Times Newspapers Ltd v the United Kingdom (47, ECHR 10), the ECtHR justified the rationale behind the relatively truncated period of time given to plaintiffs in which to bring defamation proceedings:
“The introduction of limitation periods for libel actions is intended to ensure that those who are defamed move quickly to protect their reputations in order that newspapers sued for libel are able to defend claims unhindered by the passage of time and the loss of notes and fading of memories that such passage of time inevitably entails.”
Also relevant to this issue is that at common law, each time a statement was published it gave rise to a separate cause of action, as established in Duke of Brunswick v Harmer [1849] 14 QB 185. This “multiple publication” rule meant that each individual publication of a single defamatory statement could ground an action. This raised particular issues in relation to articles published online, as in theory whenever someone accessed or read a publication on the internet, a new cause of action would accrue, notwithstanding the fact that they may already have commenced separate proceedings in respect of the same statement.
The Defamation Act 2009, however, abolished the multiple publication rule in this jurisdiction. Section 11(1) provides that “a person has one cause of action only in respect of a multiple publication.” This was considered in Dan Philpott v Irish Examiner Limited [2018] 3 IR 565, which confirmed that:
“It is the date of first publication that is to be treated as the date of publication of a multiple publication that is accessed online at different times and/or on different dates. Were this not so, online news providers would be under a never-ending duty at all hours of every day to re-visit constantly each and all of their archived articles and reports and repeatedly up-date them so that those archived articles and reports reflected any changes of relevance that might have occurred between the moment the articles initially went up on-line and each time those articles might be freshly accessed thereafter. Such an end would be a complete nonsense, commercially, practically and legally.”
The question of when a cause of action accrues – ie. when that 12 month limitation period begins – in respect of a publication on the internet was another potentially problematic issue thrown up by the internet age. Was it the date that the material was uploaded; was it the date when it was first downloaded or accessed; and did the fact that the material was permanently accessible at any time, from anywhere in the world, affect this issue?
With this in mind, the 2009 Act appeared to have clarified the matter, as Section 39 of the Act amended the Statute of Limitations Act 1957 by inserting a new Section 11(3B) to the following effect:
“the date of accrual of the cause of action shall be the date upon which the defamatory statement is first published and, where the statement is published through the medium of the internet, the date on which it is first capable of being viewed or listened to through that medium.”
This would appear to have made it crystal clear that, in respect of any material published online, a cause of action accrued on the day that the material was first published online.
This clarity, however, has been called into question by the decision of the Court of Appeal in Gilroy & Byrne v O’Leary and Google Ireland [2025] IECA 42 published on Monday 24 January 2025. Because in a wholly unexpected decision, the Court suggested that in respect of an online platform such as Google – which owns the YouTube platform – the cause of action may only accrue when the platform has been put on notice about the potentially defamatory content, regardless of how long the Plaintiff waits before providing such notice.
The Court of Appeal proceedings:
The case concerned an appeal against the High Court’s decision to refuse an application to join Google as a defendant to the proceedings pursuant to Order 15 Rule 4 of the Rules of the Superior Court. The original application, which concerned a video posted by the First-named Defendant on YouTube which the appellants say defamed them, was refused on the grounds that the application to join Google was only made four years after the video was first uploaded.
While the Plaintiffs had instituted proceedings against the First-named Defendant – the author of the statements – within a week of them being published online in June 2018, they did not engage in any correspondence with Google until more than 3 years later, and only instituted proceedings against them a further year later, in October 2022.
This would appear to suggest that an action against Google was clearly statute barred, a position which the High Court had agreed with.
The Court of Appeal, however, disagreed, and overturned the decision of the High Court.
It should properly be pointed out at this juncture that the Court of Appeal did not conclusively state that the joinder was not statute barred. Instead, it referred to the jurisprudence concerning an application under Order 15 Rule 4, which held that the application should be refused only if it was clear that the applicant was statute barred from joining the Defendant. The Court of Appeal held that there was sufficient uncertainty to merit a decision on that question being deferred until the full hearing of the matter.
The very reason advanced by the Court of Appeal for this uncertainty, however, is a highly contentious one.
The basis for this decision is summarised at paragraph 23, where they stated that it is open to debate as to whether there were
“two separate publications, first by the defendant in June 2018, and the second, more than 4 years subsequently, by Google at some time after 18th October 2022. In my view, one cannot readily dismiss that argument as wholly unsustainable.”
On this basis, the Court held that it is “arguable” that the cause of action against Google did not accrue until they received correspondence from the appellants in October 2022, and an action against them would therefore not be statute-barred.
Analysis of the decision:
It is unclear, firstly, how the Court of Appeal can justify the possibility of the same material having been published on two separate dates, four years apart, given the existence of the single publication rule, described above, which clearly appears designed to rule out that possibility.
Aside from that, the two authorities which the Court relied upon in reaching its conclusion – the Court of Appeal of England and Wales decision in Tamiz v Google Inc. [2013] 1 WLR 2151, and the regional Supreme Court of South Australia decision in Google Inc v Duffy [2017] SASCFC 130 – create further problems.
The Court of Appeal appears to have accepted submissions from the appellants that each of these cases is authority for two propositions:
- that it is only once Google becomes a publisher that the cause of action against it accrues; and
- that Google only becomes a publisher once it has been put on notice that it is hosting defamatory content.
With the greatest respect to the learned judges, there are difficulties with both of these propositions.
The issue is a particularly nuanced one, as it entails an understanding of the proposition that (a) the act of publication, and (b) the fact of being a publisher, are not the same thing at law. The correct interpretation of the Statute of Limitations in respect of publication would appear to be that the cause of action accrues with the act of publication, but not with the identification of who may or may not be legally liable as a publisher of the statement.
There is a particular problem created by defamation law, which relies on the act of “publication” to create a cause of action, and the existence of a “publisher” as a defendant, while creating a distinction between these two similar-sounding concepts. While the term “publisher” may instinctively sound like it could only be a label which attaches to the person originally responsible for the publication – the author – it actually has two other meanings.
Firstly, the common law has always held that anyone involved in the publication process of defamatory material – printers, distributors, retailers – are potentially “publishers” of that material. This is the very situation which led to the introduction of the defence of Innocent Publication for such parties. Crucially, this defence – as the very name suggests – did not mean that they were not publishers. Instead, it meant that they were “innocent” publishers – ie. not legally liable for the publication.
And to confuse matters even further, there is the secondary, more colloquial understanding of the term “publisher” – namely a commercial publisher, an organisation whose job it is to publish material, such as a news organisation, or a newspaper, magazine or book publisher. It is this, narrower, meaning of the term publisher which section 27(a) defence of Innocent Publication of the 2009 Act is clearly referring to when it states that the defence is open to parties who are “not the author, editor or publisher of the statement to which the action relates.” Because if the defence was open only to people who were not the “publisher” in the wider sense of the term – involved in any way in the publication process – then there would be no need for the defence.
So while the term “publication” refers to the performance of the act which grounds the cause of action, the term “publisher” can encompass a wide range of players, some involved in the most tangential manner in the act of publication, but all having in common the fact that they were, in some way, involved.
In order to break down this potentially vast pool of potential publishers, particularly in the age of so many online facilitators and intermediaries, the courts have tended to apportion liability by distinguishing between “primary” and “secondary” publishers. The former are the authors of the material, who bear primary liability, as well as those entities covered by the narrow definition of the term “publisher” in section 27 of the 2009 Act – the commercial publishers who bear liability for the actions of their employees, contractors and even guests.
Secondary publishers, on the other hand, are those entities whose role in, and liability for, the act of publication is diminished by the fact that they are not usually aware of the nature of the defamatory content – printers, retailers, social media platforms – unless someone brings it to their attention.
The case law cited:
Turning to the two cases used by the Court of Appeal to support their conclusions, it is admittedly regrettable that, in a passage quoted from Tamiz, the Court of appeal of England and Wales arrived at the finding that Google had “become a publisher of the material” by the act of leaving the material on its site post-notification. I believe, however, that what the Court meant was that it became liable for the material at that juncture, rather than the suggestion that it morphed from being a party with no involvement in the publication until that point, to becoming a publisher of the material upon notification. Such a change in its status flies in the face of both legal and common sense.
Becoming liable as a publisher, however – ie to bear legal liability for its position – is not the same as simply being a publisher, which can occur without any inference of whether liability attaches to that position.
This is the very reason that the defence of innocent publication is available to those parties who are, to some degree, involved in the publication process, but could not be considered to have played an active part in its creation, or be expected to have actual knowledge of its existence. It was never suggested that these parties only became involved in the publication process when someone notified them that the material may be defamatory. Instead, they became potentially liable from the minute they were involved in the publication process. They could, however, avail of the defence of innocent publication to avoid liability for their position as publishers.
It is this distinction between the fact of being a publisher, and the potential legal liability that may flow from that position, which the Court of Appeal does not appear to have properly factored into its decision.
While the decision in Tamiz may not clarify this issue, the decision of the Supreme Court of South Australia in Google v Duffy, does provide clarity. The Court of Appeal states, at paragraph 30 of its decision, that
“the appellants in these proceedings would be entitled in this jurisdiction to rely on the observations of the (Australian) Chief Justice … in support of the argument they seek to make that Google did not become a publisher of material until, at the earliest, it was notified of the offending material and of its offending nature.”
With the greatest respect to learned judges of the Court of Appeal, the Australian Chief Justice did not make these observations. He did not say that Google did not become a publisher until it was notified of the offending material. What he stated was that:
“It is Google which designs the program which authors the words of the snippet paragraph. Google’s conduct is the substantial cause of the display of the search results on the screen of the searcher’s device. The first element necessary for Google to be a publisher is therefore established. … By virtue of the extraordinary amount of material on the World Wide Web, the inordinate number of searches which are conducted and the close to infinite variations therein, it is unrealistic to attribute to Google knowledge of the contents of its paragraphs, let alone to prove that knowledge. For that reason Google should be regarded as a secondary publisher of its search results and knowledge of their defamatory contents should not be attributed to it until notice is given.
What the Australian Chief Justice suggested was that Google was clearly a publisher of some sort, due to its involvement in the publication process, but its lack of knowledge of the specific content meant that it should be categorised as secondary publisher and, for that reason, could only be liable for the material post-notification. But its involvement in the publication, and its position as some category of publisher, was clearly engaged from the moment it started to process the information in question to generate its search results, and not from the moment that it was informed of the search engine results that it produced.
Unfortunately, the Court of Appeal did not consider a more recent, and far more persuasive authority from Australia – the decision of Google LLC v Defteros [2022] HCA 27 by High Court of Australia, which is their equivalent of our Supreme Court, and therefore a higher court than that which considered Google v Duffy. This important case was not opened by either party, and not considered by the Court of Appeal, which is highly regrettable.
Because in Defteros, which concerned an almost identical issue in respect of Google’s potential liability for its search engine results, the Chief Justice held that Google did not participate in the publication process, and therefore was neither a “primary” (principal) or “subordinate” (secondary) publisher of the material. For this reason, Google did not require the defence of innocent dissemination, and the Court held that the question of whether it was on notice of the material was therefore irrelevant.
The crucial finding in Defteros was that the question of whether Google was a publisher was decided independently from any consideration of whether they were put on notice – ie. the fact of putting them on notice would be irrelevant to their classification of being a publisher. The Court decided on Google’s status as a publisher before it even considered the issue of whether it had been put on notice.
Returning to Gilroy, the outcome of the correspondence entered into by the Plaintiff with Google in 2021 was to activate the requirement under which they are obliged to “act expeditiously” pursuant to Article 6 of the Digital services Act, or “take reasonable care” pursuant to section 27 of the Defamation Act 2009, so as to avoid legal liability for the defamatory material from that moment onwards. The correspondence did not, however, “make” the recipient a publisher.
Conclusion:
Even allowing for the Court’s caveat that it was not arriving at a definitive conclusion on the point, and was simply stating that the appellants had an “arguable” case in respect of Google being properly joined as Defendants, it is difficult to accept that the issue before the Court of Appeal was anything less than clear.
The Plaintiffs clearly formed an opinion that the material in question was defamatory in June 2018, as they issued proceedings against the author within a week of it being published. They waited for a further three years, however, before contacting Google in respect of the publication. The reason for this delay is unclear. It can only be assumed that it did not occur to them that Google may have been liable as soon as the material was uploaded, or they may have believed that Google could not be liable until such time as they were notified about the content – a belief which is clearly incorrect in law.
As it stands, however, this decision of the Court of Appeal is authority for the proposition that an internet intermediary such as Google only becomes a secondary publisher of material which it hosts once it has been put on notice of such material, regardless of when the material was first published on its platform. And for that reason, a Plaintiff can wait as long as they want before instituting proceedings against the platform, on the basis that the 12-month limitation period does not begin until such time as that notice is given.
As a matter of policy, bearing in mind the intention behind the one-year limitation period, and the abolition of the multiple publication rule, this situation is highly regrettable.
Aside from policy considerations, however, the suggestion that someone becomes a publisher on the date that a Plaintiff enters into correspondence with them, rather than with regard to the established principles of defamation law, is certainly open to question.
What this decision confirms, regrettably, is that along with the questionable findings of the High Court about online platforms in Savage v Data Protection Commissioner [2018] IEHC 122, and Muwema v Facebook [2016] IEHC 519, this area of law is in need of clarity, particularly from the legislature. Unfortunately, judging by the almost total failure of the draft Defamation (Amendment) Bill 2022 to grapple with the issue of online publication, it appears that such clarity is still some way off.