The majority reasons were delivered by Abella J., on behalf of the Coram consisting of McLachlin C.J. and Binnie, LeBel, Deschamps, Fish, Abella, Charron, Rothstein and Cromwell JJ. The majority concluded that hyperlinks are, in essence, references, which are fundamentally different from other acts of “publication”.  Adopting a bright line test, the Court ruled that a defendant can only be liable if he/she uses a reference in a manner that in itself conveys defamatory meaning about the plaintiff. Only when a hyperlinker presents content from the hyperlinked material in a way that actually repeats the defamatory content, should that content be considered to be “published” by the hyperlinker. According to the Court: 

A hyperlink, by itself, should never be seen as “publication” of the content to which it refers.  When a person follows a hyperlink to a secondary source that contains defamatory words, the actual creator or poster of the defamatory words in the secondary material is the person who is publishing the libel. Only when a hyperlinker presents content from the hyperlinked material in a way that actually repeats the defamatory content, should that content be considered to be “published” by the hyperlinker.

 Where a defendant uses a reference in a manner that in itself  conveys defamatory meaning about the plaintiff, the plaintiff’s ability to vindicate his or her reputation depends on having access to a remedy against that defendant.  In this way, individuals may attract liability for hyperlinking if the manner in which they have referred to content conveys defamatory meaning; not because they have created a reference, but because, understood in context, they have actually expressed something defamatory (Collins, at paras.  7.06 to 7.08 and 8.20 to 8.21). This might be found to occur, for example, where a person places a reference in a text that repeats defamatory content from a secondary source (Carter, at para. 12).

Making reference to the existence and/or location of content by hyperlink or otherwise, without more, is not publication of that content.  Only when a hyperlinker presents content from the hyperlinked material in a way that actually repeats the defamatory content, should that content be considered to be “published” by the hyperlinker.  Such an approach promotes expression and respects the realities of the Internet, while creating little or no limitations to a plaintiff’s ability to vindicate his or her reputation. 

The Court noted that its reasons for decision dealt only with the “deep” and “shallow” hyperlinks that were in issue in the case. The Court clearly indicated that embedded or automatic links might be assessed differently.

I am aware that distinctions can be drawn between hyperlinks, such as the deep and shallow hyperlinks at issue in this case, and links that automatically display other content. The reality of the Internet means that we are dealing with the inherent and inexorable fluidity of evolving technologies.  As a result, it strikes me as unwise in these reasons to attempt to anticipate, let alone comprehensively address, the legal implications of the varieties of links that are or may become available.  Embedded or automatic links, for example, may well prove to be of consequence in future cases, but these differences were not argued in this case or addressed in the courts below, and therefore need not be addressed here.

McLachlin C.J. and Fish J. wrote joint concurring opinions. They also noted that the decision of the Court did not apply to automatic or embedded hyperlinks. They agreed substantially with the majority. However, they were of the opinion that a hyperlink should constitute publication if, read contextually, the text that includes the hyperlink constitutes adoption or endorsement of the specific content it links to.  A mere general reference to a website is not enough to find publication.

Deschamps J., in separate reasons concurring in the result, would have required two criteria to be satisfied before finding liability based on the use of a hyperlink: (1) that the defendant performed a deliberate act that made defamatory information readily available to a third party in a comprehensible form, and (2) that the defamatory matter was brought by the defendant or his agent to the knowledge and understanding of some person other than the plaintiff.

The Crookes case is not only important because of the bright line rule it developed to address when the use of hyperlinks can be defamatory. Its importance will also be measured by the expansive, policy laden, reasons of the Court and its thorough canvassing of how the innocent disseminator defence appies to ISPs.

The Court clearly took into account and was influenced by the potential impact of its decision on freedom of expression on the Internet.  According to the Court:

“The Internet cannot, in short, provide access to information without hyperlinks.  Limiting their usefulness by subjecting them to the traditional publication rule would have the effect of seriously restricting the flow of information and, as a result, freedom of expression.  The potential “chill” in how the Internet functions could be devastating, since primary article authors would unlikely want to risk liability for linking to another article over whose changeable content they have no control.  Given the core significance of the role of hyperlinking to the Internet, we risk impairing its whole functioning.  Strict application of the publication rule in these circumstances would be like trying to fit a square archaic peg into the hexagonal hole of modernity.”  

“The Internet’s capacity to disseminate information has been described by this Court as ‘one of the great innovations of the information age’ whose ‘use should be facilitated rather than discouraged’ (SOCAN, at para. 40, per Binnie J.).  Hyperlinks, in particular, are an indispensable part of its operation…” 

The Court’s strong language will have implications on future Internet cases that involve balancing important values such as regulating hate speech, the distribution of pornography over the Internet, and the regulation of unwanted commercial messages (spam such as under CASL) with the value of freedom of expression.

The reasons of the Court also clarify the liability of ISPs for disseminating defamatory materials under the innocent disseminator defence. The reasons delivered by members of the Court strongy suggest that purely passive Internet intermediaries such as Internet hosts will not be liable for publishing defamatory content merely because their systems are used to facilitate the publication of the content. However, once ISPs or other Internet interemdiaries acquire actual or constructive knowledge (knowledge of circumstances to put them on notice) that their systems are being used to disseminate defamatory content, they may be liable for publishing it unless they take steps to prevent the publication.

The majority stated the following in discussing the innocent disseminator defence:

Defendants obtained some relief from the rule’s significant breadth with the development of the “innocent dissemination” defence, which protects “those who play a secondary role in the distribution system, such as news agents, booksellers, and libraries”:  Allen M. Linden and Bruce Feldthusen, Canadian Tort Law (8th ed. 2006), at pp. 783-84; see also Society of Composers, Authors and Music Publishers of Canada v. Canadian Assn. of Internet Providers, 2004 SCC 45, [2004] 2 S.C.R. 427 (“SOCAN”), at para. 89; Philip H. Osborne, The Law of Torts (4th ed. 2011), at p. 411.  Such “subordinate distributors” may escape liability by showing that they “have no actual knowledge of an alleged libel, are aware of no circumstances to put them on notice to suspect a libel, and committed no negligence in failing to find out about the libel” (SOCAN, at para. 89; Vizetelly v. Mudie’s Select Library Ltd., [1900] 2 Q.B. 170 (C.A.), at p. 180;  Brown, at para. 7.12(6)(c); and also Sun Life Assurance Co. of Canada v. W. H. Smith & Son Ltd., [1933] All E.R. Rep. 432, at pp. 434 and 436). 

Recently, jurisprudence has emerged suggesting that some acts are so passive that they should not be held to be publication.  In Bunt v. Tilley, [2006] EWHC 407, [2006] 3 All E.R. 336 (Q.B.), considering the potential liability of an Internet service provider, the court held that in order to hold someone liable as a publisher, “[i]t is not enough that a person merely plays a passive instrumental role in the process”; there must be “knowing involvement in the process of publication of the relevant words” (para. 23 (emphasis in original); see also Metropolitan International Schools Ltd. v. Designtechnica Corp., [2009] EWHC 1765 (Q.B.)). 

Deschamps J. expressed a similar opinion stating:

 If the person who made the reference was unaware that the information referred to was defamatory, the defence of innocent dissemination may be available: Society of Composers, Authors and Music Publishers of Canada v. Canadian Assn. of Internet Providers, 2004 SCC 45, [2004] 2 S.C.R. 427 (“SOCAN”), at para. 89…

 Byrne and its progeny are consistent with the requirement that any finding of publication be grounded in a deliberate act. If a defendant was made aware (or had reason to be aware) of defamatory information over which he or she had sufficient control but decided to do nothing about it, this nonfeasance might amount to a deliberate act of approval, adoption, promotion, or ratification of the defamatory information (see, e.g., Frawley v. State of New South Wales, [2007] NSWSC 1379). The inference is not automatic, but will depend on an assessment of the totality of the circumstances…

 This requirement of a deliberate act has already been applied in the context of the Internet. In Godfrey, the defendant Internet service provider (“ISP”) had received a “posting”, which it stored on its news server. The plaintiff had notified the ISP that the posting was defamatory and requested that it be removed, but the defendant had allowed it to remain on its servers until it automatically expired ten days later. The court held that the ISP’s failure to act, once it had become aware of the defamatory information over which it had control, constituted an act of publication once an Internet subscriber had accessed the posting. In the circumstances of the case, the ISP’s failure to act amounted to a deliberate act of approval, adoption, promotion or ratification of the defamatory information.

 In Bunt v. Tilley, [2006] EWHC 407, [2006] 3 All E.R. 336 (Q.B.), the defendant ISPs were found not to be publishers because, even though they provided services, their role in the publication process was a passive one. This aspect of the decision in Bunt is a welcome development and should be incorporated into the Canadian common law. Those whose services are used to facilitate the communication of defamatory information may be shielded from responsibility for publication if they played a “passive instrumental role” in that process (para. 23)…

A case that illustrates the nature of activities engaged in by a person who plays the role of a mere conduit is Metropolitan International Schools Ltd. v. Designtechnica Corp., [2009] EWHC 1765 (Q.B.). In that case, the question was whether Google Inc. could be found to be the publisher of the search result “snippets” that were generated automatically by its search engine in response to search terms entered by users. Eady J. found that there was no publication, because Google Inc. played a passive instrumental role in facilitating the appearance of the snippets on the users’ screens (at paras. 50-51).

In the SOCAN Tariff 22 case referred to by the majority and Deschamps J., the Supreme Court used the innocent disseminator defence criteria  for determining whether an ISP is entitled to the “common carrier” defense in para. 2.4(1)(b) of the Copyright Act. They also applied it in ruling that an ISP that hosts copyright infringing content as a mere passive host is not liable for authorizing the communication of  that content to the public, but may become liable once it acquires knowledge of the infringement if it fails to take steps to prevent it.

Section 2.4(1)(b) is not a loophole but an important element of the balance struck by the statutory copyright scheme.  It finds its roots, perhaps, in the defence of innocent dissemination sometimes available to bookstores, libraries, news vendors, and the like who, generally speaking, have no actual knowledge of an alleged libel, are aware of no circumstances to put them on notice to suspect a libel, and committed no negligence in failing to find out about the libel…

In the present appeal, the Federal Court of Appeal stated that, in the case of host servers, “an implicit authorization to communicate infringing material might be inferred from their failure to remove it after they have been advised of its presence on the server and had a reasonable opportunity to take it down” (para. 160).  Reference was made to Apple Computer Inc. v. Mackintosh Computers Ltd., reflex, [1987] 1 F.C. 173, aff’d 1990 CanLII 119 (SCC), [1990] 2 S.C.R. 209, at pp. 211 and 208, citing C.B.S. Inc. v. Ames Records & Tapes Ltd., [1982] 1 Ch. 91, at p. 110, i.e., an Internet Service Provider may attract liability for authorization because “. . . indifference, exhibited by acts of commission or omission, may reach a degree from which authorisation or permission may be inferred.  It is a question of fact in each case . . . .”  See also Godfrey v. Demon Internet Ltd., [1999] 4 All E.R. 342 (Q.B.)…

I agree that notice of infringing content, and a failure to respond by “taking it down” may in some circumstances lead to a finding of “authorization”.  However, that is not the issue before us.  Much would depend on the specific circumstances… 

In the Crookes case both the majority and Deschamps J., used the SOCAN case as an authority for interpreting and explaining the scope of the innocent disseminator  defense. The Court thus applied the same innocent disseminator principles for assessing the liability of ISPs for copyright and defamation purposes when they acquire actual or construcitve knowledge that their systems are being used in a tortious way by third parties.

Deschamps J., would also have applied these principes in determining a person’s liability for hyperinking.

The Court has in the past alluded to the existence of an “innocent dissemination” defence: SOCAN. In that case, this defence was described as follows (at para. 89):

[T]he defence of innocent dissemination [is] sometimes available to bookstores, libraries, news vendors, and the like who, generally speaking, have no actual knowledge of an alleged libel, are aware of no circumstances to put them on notice to suspect a libel, and committed no negligence in failing to find out about the libel.

In my view, the innocent dissemination defence should be made available to hyperlinkers, provided that the criteria mentioned in SOCAN are satisfied. There are good reasons for so holding. Hyperlinkers have not created or developed the information to which they link. Moreover, although the creation of a hyperlink may often be a deliberate act that makes the creator’s role more than a passive one, the hyperlinker often has little control, if any, over the linked information. This last point is significant given the dynamic nature of the Internet. A hyperlink may originally be created to link to information that is not defamatory, but the linked information may subsequently be altered so as to become defamatory without the hyperlinker’s knowledge. In such circumstances, provided that the criteria mentioned in SOCAN are satisfied, the innocent dissemination defence should operate until the moment the hyperlinker becomes aware that he or she is now linking to allegedly defamatory information. If this defence were widely available, it should dissuade overeager

What is interesting to note about the Court’s discussion of the applicability of the innocent disseminator defense to ISP liability, is the Court’s endorsement of the principle that ISPs have to take active measures to remove or disable access to defamatory content from their systems to prevent further dissemination to avoid being liable for defamation. It reached this conclusion after considering its own copyright decision in the Tariff 22 case and after considering intervention briefs from well known advocates for free speech and human rights such as Canadian Civil Liberties Association, Samuelson-Glushko Canadian Internet Policy and Public Interest Clinic, and British Columbia Civil Liberties Association.

The Court’s decision in this regard is not surprising as in the Tariff 22 case the Court recommended that Parliament implement a notice and takedown system to deal with copyright infringing content. It stated that an effective process for dealing with infringing content on host systems “would be the enactment by Parliament of a statutory ‘notice and take down’ procedure as has been done in the European Community and the United States.”

On March 22, 2010 – before the federal election was called- TELUS, Bell and Rogers appeared before the Special Legislative Committee studying Bill C-32. The ISPs continued to endorse notice and notice asking that this process be formalized in C-32. Studies around the world have shown that notice and notice by itself-without any real threat of a sanction- is not the most effective way of reducing online peer-to- peer file sharing. TELUS, Bell and Rogers which have, to some extent, been voluntarily passing on notices from rights holders for about decade, were asked by the Parliamentary Committee whether notice and notice was effective. As it turns out, after a decade of experience, it became clear from their testimony that the ISPs could not show to what extent notice and notice (without any real threat of a sanction) deters online file sharing.

The MPs on the Committee were keenly aware that the ISPs had advocated for notice and notice and rejected any form of graduated response to provide further assistance in reducing online file sharing. So, the MPs on the Committee, lead by Mr. Marc Garneau and Mr Pablo Rodriguez, asked the ISPs to provide evidence that notice and notice without the possibility of sanctions is effective. Mr. Garneau squarely challenged Canada’s leading ISPs to provide data to support their positions:

Mr. Marc Garneau: The question that comes up – because I hear from a lot of groups – is whether notice and notice is effective. There are consumer groups and others that say, yes it is effective, it’s working, and we don’t need to ramp up to notice and takedown, or some other hybrid version of a more proactive approach, if you like. What’s missing for me as an engineer is the data. I believe you have that data because you have been providing notice on a voluntary basis to people who are stepping over the line. I’m really interested in it, because some people say that notice and notice is not discouraging those who know it has no teeth while other people are saying, yes as soon as you get a notice and notice in your home from your ISP provider, it sends a chill through you and you realize you’d better not do it again.

You have the data to tell us whether repeat offenders are in large numbers, or whether 95% of people who have received their first notice cease and desist. I would like to hear from each of you if you have that data. If not I’m going to ask you to present it to this committee, because we really do need that data. If you have it, it would be great to hear from each of you, perhaps starting in the way you spoke about whether it is actually effectively working, and I’d like to see some statistics to support that.

The ISPs, who had claimed to be engaging in notice and notice for almost 10 years, did not provide the Committee with any study, survey or independent evidence whatsoever to establish the effectiveness of notice and notice.

Mr. McTaggart who appeared for TELUS told the Committee that TELUS had no data whatsoever to establish the effectiveness of  notice and notice. TELUS, which according to Mr McTaggart has been forwarding notices to its subscribers for 9 years, had only “anecdotal” stories.

Mr. Craig McTaggart: For TELUS’ part my answer is a very short one. We don’t actually have that data. We only forward the notices. We don’t retain any further data about how many messages were sent to a certain customer in a certain period, because we have no business reason to retain that data. We only have the anecdotal stories, as you referred to, that often a household receives a notice, the parent reads it, doesn’t know anything about what it’s about, talks to a child, and gives the strong message not to do it anymore…

Mr. Marc Garneau: So if John Smith gets a notice from you today or in the past, we and you don’t know whether in fact you’ve sent John Smith 25 notices over the course of time because John Smith is clearly ignoring the notice and notice.

Mr. Craig McTaggart: At TELUS we don’t know that.

Ms. Suzanne Morin, who appeared for Bell, said that Bell was in the same position as TELUS. She said that Bell “thinks” it’s effective and has an educational impact … “But at the end of the day it still needs something and so we’re waiting for legislation to be passed so that users know specifically that downloading illegally is not to be tolerated in Canada.”

Ms. Pam Dinsmore appeared for Rogers. She presented some data on Rogers’ experience. Here is what she said:

Ms. Pam Dinsmore: I would like to tell you what happens at Rogers. We can actually track the number of notices that go to an individual household. What we can’t track is how many Sony notices went to a particular household. In terms of the actual processing and sending on of a notice, we do have, at a very high level, those numbers.

Let me put it in perspective for you. On a Rogers customer base, for argument’s sake, of 1.5 million Internet customers, about 5% of that customer base will receive a notice. As I said, in 2010 we processed 207,000 notices.

If I go down the pipe to the next level, of that group, of that 5% of our customer base getting a notice, which is in the area of about 70,000 customers, the next number of households to get a second notice will be somewhere in the area of—, I don’t know—, 20,000 notices. So about a third of those who got the first notice will actually get a second notice.

As I get to the third level, I’m down a third again. So as I go down the list, the number of notices drops in accordance with the times the notices are received.  In our view, the notice and notice routine is effective at discouraging those people who are alleged to have infringed — only alleged to have infringed — from infringing again. We think it does put the fear of God into them and it is effective in doing that.

That’s based on the information that we are able to track.

Anti-copyright advocate Michael Geist argues that these statistics “provides solid evidence that notice-and-notice is effective in countering repeat infringement”. He concludes that “the Rogers data suggests that 67% of recipients (which is already only five percent of subscribers) do not repeat infringe after receiving a notice and 89% cease allegedly infringing activity after a second notice. Within two notices, about 99% of Rogers subscribers are not receiving infringement notifications.”  But, the data does not establish this. What can we conclude from Rogers’ data?

First, the data does not tell us anything about what percentage of Rogers’ customers actually use its systems to illegally share files. According to Rogers, 5% of its customers got notices of claimed infringement in 2010. But, Rogers admitted it has no idea what percentage of its customers actually use its systems to infringe. This information was disclosed in response to questions from MP Pablo Rodriguez:

Mr. Pablo Rodriguez: Okay. Ms. Dinsmore, you said that roughly 5% of your clients get a notice. Does that mean that 95% of your clients comply with all the regulations?

Ms. Pam Dinsmore: They simply don’t get notices that are directed towards their households.

Mr. Pablo Rodriguez: Do you think that 5% is the actual number of people who download things illegally, or are there many who get away with it, without receiving anything?

Ms. Pam Dinsmore: I have no way of knowing that.

Second, Rogers’ data suggests there is an enormous file sharing problem occurring through the use of the ISPs’ networks. Rogers’ data suggests that at least 70,000 of its account holders were engaged in illegal downloading in 2010. The number could be far higher because Rogers says it has “no way of knowing” how many of its account holders are using its systems to infringe. Worse there appear to be at least 20,000 households who recommit infringing acts after being caught.

The existence of an enormous problem was confirmed by testimony from Ms. Suzanne Morin from Bell. She testified that in 2010 Bell received over a million notices of claimed infringement from copyright holders. Bell received so many that it was not able to process them all. To do so Bell “would have to fill a whole floor with individuals in order to process them all.”

Third, there are repeated copyright infringements by Rogers’ subscribers. Rogers’ testimony is that some of its account holders are getting 6 or more notices of claimed infringement. Rogers appears to be saying that it issued 70,000 first time notices (about 5% of Rogers 1.5 million customers which is 75,000 not 70,000 customers); 20,000 second notices (less than a third of 70,000, its 28.5%) and a reducing similar number of notices as Rogers continues to send notices to alleged repeat infringers. Based on the testimony, there are notices being sent at least 6 times to around 133 account holders. The table below extrapolates Rogers’ numbers to the 6th notice. 

Fourth, Rogers’ evidence doesn’t tells us the effectiveness of notice and notice. What the Parliamentary Committee wanted to know was to what extent individuals who received a notice stopped online file sharing after getting the notice. But, Rogers’ data doesn’t answer this critical question.

It is conceivable that receiving a notice deters 2 customers out of 3 to stop infringing. But the figures don’t prove this. We cannot assume that 100% of account holders who get a first notice and who keep infringing are detected by rights holders and sent a second notice. Surely a large number of infringers don’t get second or follow on notices simply because the probability of getting caught is not 100%. If the probability for an infringer to get caught and warned is less than 100% one would expect a declining number of notices to the same account holder.

So did the number of second and follow on notices decline because the notices had an effect on illegal activity or because the continued file sharing was not detected, for some other reason, or for some combination of reasons? The notices may have had some effect as it is logical to conclude that some people would be discouraged from further illegal file sharing when receiving a notice of claimed infringement. But, we cannot ascertain how much of an effect it had from the disclosed data. It seems fair to assume that if Rogers does not know how many of its customers that infringe for the first time are sent notices of claimed infringement, Rogers also does not know how many of its customers who received a first notice continue to infringe but are not caught and so are not sent a second or follow on notice. Accordingly, there is no basis in the data disclosed by Rogers to draw any conclusions about how effective the notice and notice routine is at discouraging those people who are alleged to have infringed from infringing again.

To determine the effectiveness of Rogers’ notices, one would need to follow the perpetrating households through time to determine the extend to which receipt of a notice has actually lead to a reduction in online file sharing. One would also want to go further and focus on the question of the relevant counterfactual[1].  One would want to compare the effect of a notice against the results under a different regime. Rogers’ data doesn’t answer  the real questions the Committee was concerned to learn about, namely: (a) has notice and notice been effective on changing behaviour compared to no notice at all, and (b) would notice and notice with a real potential sanction or threat of a real sanction generate a better effect than a notice and notice not backed-up by a realistic threat of a sanction  to deal with repeat infringers?

On counterfactual (a), if there was no notice and notice regime, one might have recorded the account holders’ behaviour and seen some decline in repeat infringements. In such event, the notices would not be the only circumstance impacting on the flow of notices.

On counterfactual (b) Rogers’ data does not reveal whether the threat of a real sanction being carried out would result in less recidivism. If the counterfactual is a regime with a real threat of a sanction being carried out– simple economics predicts people make cost/benefit decisions and that a real threat would change behavior, but that no real threat is unlikely to do so to the same degree. (Having a speed limit that is never enforced or which has no sanction even if enforced will not have as much effect on reducing speeding than one that has a sanction that is enforced.)

Rogers’ data also does not tell us which category of its account holders are causing the most damages. It may be the recidivisits who are ignoring the notices are causing a disproportionate amount of the harm.

In the end, after a decade of notice and notice, the three ISPs were unable to provide the Committee with any reliable data or studies to back up their claims that notice and notice is a good policy choice because it is an effective deterrent or at least as good a deterrent as other means. MP Pablo Rodriguez summarized his interpretation of the ISPs’ testimony at the Committee as follows:

Mr. Pablo Rodriguez: That’s the crux of the problem. You are telling us that notice and notice is enough, but, on the one hand, you have no data, and, on the other hand, you are not able to tell me whether the 5% is representative or not…

Yet you insist that notice and notice is enough. I am not an engineer, but I would like to have data to rely on, and we don’t.

Bill C-32 is behind us. We now need to start thinking about what policies will work best to stem online file sharing to help foster the legal market for digital products. In doing this we need to understand that the threat that rights holders will sue their customers in Canada is not and has not been a deterrent to illegal file sharing. Further, one cannot expect that simply sending notices to account holders when there is no realistic expectation of a penalty to be imposed by someone will be as effective a means to deter illegal online file sharing as one that has a substantial deterrent. Reducing statutory damages for either non-commercial or private purposes would also send a signal that there will be fewer sanctions for engaging in illegal file sharing. From an economic standpoint, the deterrence relies on the expectancy of the sanction.[2]

The Full Court of Australia in the iiNet case recently unanimously found that notices are likely to be effective in reducing infringements when coupled with potential action such as suspension or termination of accounts of customers that fail to heed such notices. The judges also expressed the unanimous opinion that taking steps to adopt and implement a graduated response system would be reasonable and likely effective to prevent repeat infringements by its subscribers.

Canadian ISPs have pushed for notice and notice to be in the previous copyright reform amendments, Bills C-60, C-61, and C-32. However, after a decade of notice and notice, the ISPs have not produced evidence of the effectiveness of notice and notice or its relative effectiveness against other options including some form of graduated response such as what countries like France, UK, and South Korea have implemented. It is now time to examine alternatives that would have the intended effects which presumably everyone wants to achieve.

*Note blog amended April 4, 2011.

[2] Olivier Bomsel and Heritiana Ranaivoson entitled “Decreasing Copyright Enforcement Costs: The Scope of a Graduated Response”, Review of Economic Research on Copyright Issues, 2009, vol. 6(2), pp.13-29.

A main issue in the appeal was whether iiNet had an obligation to have and implement a policy to transmit notices of claimed infringement to its subscribers and to terminate the accounts of repeat infringers who used its systems and services to engage in file sharing. The trial judge had held that iiNet was not liable for infringement for refusing to do so. The appeal was rejected by a 2 to 1 majority, with Emmett and Nicholas JJ concurring in separate opinions that the appeal should be dismissed, and with Jagot J, who would have allowed the appeal.

All three judges were of the opinion that iiNet could have established and implemented a policy to forward notices of claimed copyright infringement to subscribers that could have included suspending and terminating the accounts of repeat infringers to avoid being liable for authorizing copyright infringement under Australia’s copyright laws. All three judges also held that iiNet’s obligation to act was premised on copyright holders providing cogent evidence of the alleged primary acts of infringement which took place using its services.

The main disagreement between the judges was whether iiNet had been provided with sufficient information to trigger the obligation to act. Jagot J. was of the opinion that iiNet was given everything required and in any event would not have acted irrespective of the information provided to it. Emmett and Nicholas JJ expressed the opinion that sufficient information had not been supplied. Nicholas J found that the fact that iiNet may not have acted irrespective of the information was “not to the point”.

Accordingly, while iiNet was exonerated on the facts, the opinions rendered in the case leaves it open for copyright holders in Australia to hold ISPs liable for authorizing infringement unless they take available steps to deal with infringements by their subscribers.

The decision is an important one for ISPs and rights holders in Australia. However, the decision also has global implications for policy makers that are concerned with finding practical and efficacious means of reducing the scourge of illegal online distribution of copyright content. I say this because in the course of giving reasons for decision, the judges of Australia’s appeal court canvassed the effectiveness of sending out notices of claimed infringement with warnings of potential account terminations and found they would be effective in reducing online file sharing. They also reviewed the reasonableness and practicality of requiring ISPs to implement a graduated response system including sending out these types of notices and adopting and implementing a process to terminate accounts of repeat infringers. They found such a process reasonable and workable.

As noted above, the main issue in the iiNet case was whether iiNet was liable for the authorizing infringement of its subscribers who used its systems and services to engage in illicit peer-to-peer file sharing. To make this determination the court canvassed a series of issues including whether its subscribers were liable for copyright infringement when they used the BitTorrent protocol to share movies and television programming and whether failing to implement a graduated response system rendered iiNet liable for authorizing infringement.

Liability of subscribers for copyright infringement

The first issue in the case was the extent of infringement by iiNet subscribers using BitTorrent to engage in file sharing of movies and TV programming. While iiNet had admitted that users had infringed copyright, it claimed that they only ever did so once and that there were no further infringements for iiNet to prevent. The trial judge had agreed. The trial judge had found that users of BitTorrent had infringed copyright, however the parties disagreed on the extent of infringements taking place and this was potentially relevant to the obligations that iiNet would be under to act on the infringements. The trial judge had also found that users of BitTorrent only ever infringed copyright once, the first time the file was shared with other users, irrespective of the duration the file was available to other BitTorrent users.

All of the appeal judges reversed the finding and found that the uploaders were liable for multiple infringements of copyright.

Australia has implemented the WIPO Internet Treaties. Accordingly, unlike Canada, its copyright laws include a right of making available a work to the public. All judges of the court accepted that when an individual uploads a file to make it available for sharing over a BitTorrent network, the person is liable for infringement under the making available right. The court also held that each time an individual connected to the internet, a separate act of making the work available to the public occurred.  All of the judges also agreed that the electronic transmissions made by subscribers as part of a BitTorrent swarm involved transmissions that were “to the public”. However, the judges disagreed as to whether the evidence before the court established that files or substantial parts of files were communicated to the public when transmitted by the BitTorrent protocol.

Authorization

The main issue in the case was whether iiNet was liable for the illegal file sharing of its customers under the theory of authorization. Section 101(1) of Australia’s copyright law establishes liability for authorization in Australia:

101 (1)  Subject to this Act, a copyright subsisting by virtue of this Part is infringed by a person who, not being the owner of the copyright, and without the licence of the owner of the copyright, does in Australia, or authorizes the doing in Australia of, any act comprised in the copyright.

The Australian Act was amended in 2000 by inserting into s. 101 a new subsection 1(A) to provide a practical enforcement regime for copyright owners and to promote access to copyright material online.  Universal Music Australia Pty Ltd. v. Sharman License Holdings Ltd., [2005] FCA 1242 (5 September 2005); Universal Music Pty Ltd. v. Cooper, [2005] FCA 972 (Aust. F.C.). The provision states that in deciding whether a person has authorized the doing in Australia of any act comprised in the copyright, the matters that must be taken into account by the Court include:

(a) the extent (if any) of the person’s power to prevent the doing of the act concerned;

(b) the nature of any relationship existing between the person and the person who did the act concerned;

(c) whether the person took any other reasonable steps to prevent or avoid the doing of the act, including whether the person complied with any relevant industry codes of practice.

In the course of analyzing the above factors, and in particular factors (a) and (b), the judges canvassed whether iiNet had the power to prevent illegal file sharing by its subscribers and could take reasonable steps to prevent or avoid these infringements by implementing a graduated response system. The judges of the court expressed the unanimous opinion that iiNet had such a power and that taking steps to adopt and implement a graduated response system would be reasonable and likely effective to prevent repeat infringements by its subscribers – even though two judges found that on more limited grounds it was not unreasonable for iiNet not to act in this case (based on the knowledge iiNet had before the case had been filed).

Effectiveness of a notice and notice regime coupled with a threat of account termination

iiNet had argued in the appeal that it should not be required to send its subscribers notices of claimed infringement with warnings of potential account terminations because such notices would not be effective in stemming online file sharing. Some Canadian ISPs have had a de facto notice and notice system in Canada for years. Despite this, there is no evidence that it changes people’s behavior to stop illicit file sharing and purchase creative products from legitimate services. However, as I and others have pointed out elsewhere, research by our trading partners shows that a notice with a threat of some sanction will operate as an effective deterrent. Opinions rendered by the iiNet court provides added support for the that fact that notices are likely to be effective in reducing infringements when coupled with potential action such as suspension or termination of accounts of customers that fail to heed such notices.

Emmett J. expressed this opinion as follows:

There was no finding by the primary judge that infringing acts engaged in by iiNet users were engaged in, in circumstances where the customer whose account was being used knew, or had reason to suspect, that infringement was involved. It may well be that an iiNet user would not know that, by downloading the Films by means of the BitTorrent System, there was infringement. It may be that an iiNet customer was unaware that the service provided to that customer was being used for infringement. A warning, coupled with reference to the entitlement of iiNet to terminate the service, may well be sufficient to persuade a customer to desist from engaging in such conduct or to take steps to ensure all users of that customer’s service desist. It is one thing to be aware of a prohibition on infringement of copyright. It is another thing to be aware that specific acts engaged in actually constitute infringement.

It may be possible that some, or indeed many, customers would disregard a warning given by iiNet. However, it does not follow, from that possibility, that all customers would disregard a warning or even that most customers would do so. Even if many customers may disregard a warning, compliance by others would prevent further infringement involving the accounts of those customers from taking place. Every time a user of the internet leaves a BitTorrent swarm, the supply of copyright material made available online within the swarm would be reduced and the opportunity for other internet users to infringe copyright would therefore also be reduced.

Giving clearly worded and accurate warning to a customer whose account was thought to be involved in acts of infringement is a reasonable step that could have been taken by iiNet, regardless of whether the warning would have been heeded by every customer. There is no basis for concluding that all, or even a substantial proportion, of iiNet’s customers would permit their accounts to continue to be used for acts of infringement if they were properly warned that the acts constituted infringement and their attention was drawn to their contractual obligations and iiNet’s contractual right to enforce those obligations by suspension or termination.

Nicholas J expressed a similar opinion:

The primary judge seems to have come to the conclusion that the giving of a warning could not be a reasonable step for the purposes of s 101(1A)(c) because the mere giving of a warning is not a power to prevent. I would accept that the ability to give a warning is not in itself a power to prevent for the purpose of s 101(1A)(a). However, s 101(1A)(c) is concerned with a different matter namely, what “reasonable steps” the alleged authoriser might have taken to prevent or avoid any of the relevant acts of copyright infringement.

I am satisfied that the primary judge erred in holding that the giving of a warning could not be a reasonable step that might be taken to prevent or avoid the relevant acts of copyright infringement. If the ability to give a warning is supported by a power to terminate or suspend a subscriber’s account then there can be no doubt that the giving of a warning is capable of amounting to a reasonable step for the purpose of s 101(1A)(c).

So did Jagot J:

It is also difficult to accept that most people when notified of copyright infringement by them or a person using their service would simply ignore the notice unless threatened with termination. Not all people would be aware of the risk of acting unlawfully in downloading films and television shows from the internet. Not all people would be aware that their downloading activities can be monitored both by third parties (such as copyright owners using software like the DtecNet agent) and internet service providers. When confronted by a mere notice or warning with evidence both of copyright infringement (either by them or another person they permitted to use their internet service) and the ease with which it can be detected by third parties, it is difficult to accept that there would be no deterrent effect whatsoever. To the contrary, it could readily be assumed that many people on receipt of a mere notice or warning would be deterred from future infringements irrespective of termination. The fact is that by receipt of a mere notice or warning people would realise that activities they might have thought innocent or at least undetectable were in fact unlawful and open to scrutiny.

The judges also commented on the reasonableness of an ISP with a contractual right to terminate an account doing no more than merely posting a policy on a website warning customers about this policy. They did not view taking these steps as being sufficient to constitute a reasonable response to the infringements. Emmett J, for example, stated the following:

The language of s 101(1A)(c) requires an enquiry as to any steps that were taken by iiNet and whether there were reasonable steps that were not taken. The Infringement Notices specifically invited iiNet to cancel, suspend or restrict the accounts of the customers identified in the spreadsheets enclosed with them. Merely having a contractual provision relating to copyright infringement and drawing attention to the contractual provision on its website is hardly an effective step to prevent infringement, in circumstances where iiNet does not enforce the contractual provision.

Reasonableness  of account termination

The judges of the Full Court also canvassed in detail the reasonableness and practicality of iiNet adopting and implementing a policy to terminate the accounts of repeat infringers. The primary judge had held that implementing such a system would not be reasonable, apparently under any circumstances. Critics of effective enforcement of copyright like Prof. Geist have relied on this holding to argue against requiring ISPs to ever take steps to suspend or terminate accounts of repeat infringers. However, the Full Court expressly disagreed with the primary judge on this issue and delivered a strong endorsement of such steps in order to prevent continuation of copyright infringement.

Emmett J. pointed out that iiNet terminated user accounts for other reasons. He also pointed out that terminating a user’s account where the user uses the account in violation of contractual terms of service to infringe copyright is reasonable and was exactly the step contemplated by the safe harbour provisions of the copyright law.

Even where a service provider such as iiNet has the benefit of the Safe Harbour Provisions, the Court is specifically empowered, under s 116AG(3)(b), to order termination of a specified account. It can hardly be concluded, therefore, that termination was, per se, unreasonable. Rather, the Copyright Act itself contemplates such a step. Accordingly, it must be regarded as a reasonable step, at least in some circumstances, including circumstances involving repeat infringements, to terminate or suspend an account of a customer.

Thus, iiNet had the capacity to control the use of its services by its customers and to take steps to prevent acts of infringement by the use of services provided to them. In some circumstances, iiNet did in fact exercise control over the provision of its services. The presence of such provisions amounted to a degree of power to prevent further acts of infringement that was significant, for the purposes of s 101(1A)(a). That power arose from the relationship between iiNet and its customers under the customer relationship agreements, the nature of which is significant for the purposes of s 101(1A)(b).

There is no reason to conclude that sending warnings would be unreasonable, given that iiNet’s business routinely involved sending warnings to users in a variety of circumstances, such as when fees were overdue. There is no reason to view a temporary suspension as unreasonable, particularly when iiNet had the right to terminate the service altogether. The period of a temporary suspension would be a matter for consideration in any given case. A very long suspension for a very minor infringement may be unreasonable. However, that is not a reason for concluding that no suspension could ever be reasonable…

iiNet says, nevertheless, that termination of an account was not a reasonable exercise of a power to prevent because it would involve disconnection of the customer from the internet. Disconnection would not, in itself, prevent infringement because of the possibility of the customer obtaining access to the internet through another carriage service provider. On the other hand, disconnection would do much more, in that the whole relationship between iiNet and its customer would be ended, thereby foreclosing all internet activity by that customer by means of iiNet’s service.

However, it is difficult to see why that consequence is unreasonable, if the customer, having been warned that the service being provided to the customer by iiNet was being used to engage in infringing acts, that to do so was a breach of contract and that continuing to do so may result in termination, nevertheless chooses to continue to permit the iiNet service to be used to engage in infringing acts. It may be unreasonable to terminate an account without warning. However, that is not the question.

Jagot J reached the same conclusion rejecting the contention that adopting and implementing a graduated response system was unreasonable.

Despite this, iiNet adopted and maintained thereafter a clearly stated position that the AFACT notices did not identify any actual infringing user and, in any event, that iiNet had no obligation to do anything in response. This first aspect of iiNet’s position was adopted despite iiNet knowing not only that it could match IP addresses to customer accounts but also that its policies for dealing with other forms of unlawful or abusive activity assumed that it would do so and thereby hold customers responsible for all use on their accounts…

Another reason I take a different view is that iiNet’s position on the use of its service in a manner infringing copyright was not only inconsistent with its approach to other management issues (such as customers not paying bills or exceeding their quota), but also inconsistent with its approach to other types of internet abuse (such as spamming). The unavoidable inference is that when its own interests were at stake, iiNet exhibited no hesitation in: – (i) using IP addresses to identify the relevant customer accounts, (ii) treating the customer as responsible for all use of the iiNet service on the customer’s account, and (iii) promulgating a regime of warnings, suspension and termination (albeit discretionary) of the customer’s account. However, when its own interests were either not at stake or, at worst, might have been adversely affected by taking action, iiNet adopted a contrary position…

…it is feasible for an internet service provider to respond to the receipt of credible evidence of copyright infringements by users of its service…there is no particular difficulty with an internet service provider holding a customer responsible for the use of that customer’s service (consistent, in this case, with iiNet’s CRA and other policies about internet use)…

An internet service provider is confronted by precisely the same problems when dealing with spamming and other forms of internet abuse. iiNet has been able to formulate (and presumably implement) its policy for dealing with network abuse…

I do not see the expense and complexity of implementing a policy or scheme of warnings and suspension or termination as insuperable difficulties rendering the taking of such steps as unreasonable. The trial judge considered that the complexity and expense of such a scheme “manifestly militates against the conclusion that such scheme is a relevant power to prevent” (at [435]). For the reasons given, I consider that this conclusion does not accord with either the factual reality (that iiNet was capable of implementing equivalent schemes to deal with other issues such as network abuse and spamming) or the legislative scheme (which, by the safe harbour provisions, contemplates schemes of this very kind at least for repeat offenders).

Nicholas J. reached a similar conclusion.

Nor do I think that the difficulties involved in establishing a system for giving warnings and, if necessary, termination or suspension of accounts, were likely to be as great as Mr Malone’s evidence might suggest. It is true that the respondent, if it was to take the step of issuing warnings in particular cases, would need to decide when it would be appropriate to do so. To that end, the respondent would need to decide, among other things, whether the available material was sufficient to satisfy itself that it was appropriate for a warning to be issued or for an account to be terminated. It is also true that the respondent would need to decide how many warnings should be given (unless it decided that no warnings need be given) before terminating or suspending service to a particular subscriber. These and like questions involve matters of judgment and degree. As I have already acknowledged, the decision as to whether or not to terminate a particular account may not be a simple one. But I do not accept that the adoption of some system providing for the issuing of warnings followed by termination or suspension is not a reasonable step which the respondent could have taken for the purpose of preventing or avoiding copyright infringement by users of its network.

Nicholas J, also expressed the opinion that in the absence of a industry code of conduct that iiNet had considerable flexibility as to how to design an effective and fair graduated response process.

Nevertheless, in the absence of applicable regulations or access codes which might guide an ISP’s decision making in relation to such questions, it seems to me that an ISP should be given considerable latitude when working out the detail of such a system. It is always possible to argue that a system for the issue of warnings and termination could be tougher than it is. But it would be difficult to criticise an ISP on that account if it acted in good faith to devise and implement a system that involved taking such steps against subscribers who the ISP was satisfied had used (or permitted others to use) its facilities for the purpose of committing flagrant acts of copyright infringement.

The knowledge of infringement requirement

All of the judges accepted that the reasonableness of iiNet being required to take steps to forward notices of claimed infringement and to terminate accounts depended on its having reliable evidence that an account was being used to infringe copyright. On the facts of the case, two out of the three judges expressed the opinion that the information given to iiNet was not adequate to meet the required standard. In the final analysis that was the chief (possibly sole) reason why the appeal was dismissed.

Nicholas J. stated that to trigger a graduated response by an ISP, a notice of claimed infringement sent to the ISP must contain an explanation of how the information about the subscriber infringement was collected. He also suggested that a notice addressed to an ISP should contain a statement verifying the accuracy of the data or the reliability of the methods used to collect it. Emmett J, stated that what is required is

unequivocal and cogent evidence of the alleged primary acts of infringement by use of the iiNet service in question. Mere assertion by an entity such as AFACT, with whatever particulars of the assertion that may be provided, would not, of itself, constitute unequivocal and cogent evidence of the doing of acts of infringement. Information as to the way in which the material supporting the allegations was derived, that was adequate to enable iiNet to verify the accuracy of the allegations, may suffice. Verification on oath as to the precise steps that were adopted in order to obtain or discern the relevant information may suffice but may not be necessary.

Emmett J, would also have imposed two additional requirements on copyright owners before requiring them to voluntarily adopt and implement a graduate response process. First, that the copyright owners had undertaken “to reimburse iiNet for the reasonable cost of verifying the particulars of the primary acts of infringement alleged and of establishing and maintaining a regime to monitor the use of the iiNet service to determine whether further acts of infringements occur. Second, that the copyright owners had also undertaken to “indemnify iiNet in respect of any liability reasonably incurred by iiNet as a consequence of mistakenly suspending or terminating a service on the basis of allegations made by the Copyright Owner.” Emmett J, was the only judge to have added these two requirements.

The issue of whether iiNet had sufficient knowledge of the infringements to act was a hotly disputed issue. However while the trial judge had held that even though iiNet did not have a complete understanding it had decided from the second notice that it would never act on the notices irrespective of the information supplied to it.  Despite the fact that this finding was unchallenged, the majority judges (Emmett J and Nicholas J) did not consider the decision of the ISP not to act was determinative.  In a strong dissent, Jagot J reached the opposite conclusion.

Some reflections on the iiNet case

Online file sharing is a major problem in Canada. Recent research in Canada and elsewhere demonstrates the adverse impacts and economic losses that result from illicit online file sharing in the film and music industries.

The findings of the Full Court in the iiNet case show that graduated response processes can be effective, reasonable, and fair mechanisms to reduce illegal file sharing. In fact, although not referred to in the decision, last year the Irish High Court in EMI Records & Ors -v- Eircom Ltd, [2010] IEHC 108 reach a similar conclusion. The Irish court held that a graduated response solution to illegal file sharing involving “detection, notification and termination” “is viable and proportionate”. Other academic research supports this view as well. See, Prof.Bomsel Decreasing Copyright Enforcement Costs: The Scope of a Graduated Response, Prof.Strowel, Internet Piracy as a Wake-up Call for Copyright Law Makers—Is the ‘‘Graduated Response’’ a Good Reply?, See also, Barry Sookman and Dan Glover Graduated response and copyright: an idea that is right for the times.

The iiNet decision provides useful guidance on how Bill C-32 could be amended to better achieve the stated objectives of the Government and the opposition parties. Bill C-32 already contains a notice and notice regime which requires online service providers to forward notices of claimed infringement to end users. The current proposed notice and notice process lacks efficacy, however, because the notices do not carry any threat of sanction for repeat infringers. The iiNet decision suggests that an additional requirement to mandate online service providers to adopt and implement a policy to terminate the accounts of repeat infringers would be an efficacious and reasonable way to reduce online file sharing in Canada.

* Links updated Mar. 13, 2011

In the course of reaching this decision, the court made a number of important rulings about the liability arising from the use of BitTorrent networks including the following:

• Seeders and peers that make music available for sharing are infringers under Australia’s making available right.
• The transmission of copyright files as part of a BitTorrent swarm constitutes a transmission (communication) to the public by participants in the stream and is infringing.
• A peer that downloads a copyright file over an unauthorized BitTorrent network infringes the reproduction right of the copyright owner.
• There can be liability for the transmission over a BitTorrent network even if some part of the transmission occurs outside of Australia.

The court also strongly suggested that operators of BitTorrent sites would be liable for copyright infringement for having authorized infringement.

As noted above, the court held that in the circumstances of the case, an ISP that simply provides internet connectivity to subscribers does not by that act alone authorize the subscriber’s infringing activity.

The result of the case has left many Australians thinking that some solution to stemming illegal file sharing is necessary. One solution being discussed is an industry wide code of practice between ISPs and content owners. The establishment of a code would have legal effect as it is an express factor in the Australian Copyright Act for determining whether a person authorizes an infringement.

Another solution being discussed is a state sanctioned graduated response system. The Sydney Morning Herald just published an article, Markets and laws failing in face of net piracy, pointing out that there was a real problem to be addressed and that something needed to be done. The article put it this way:

“In his ruling last week, Justice Dennis Cowdroy noted this was the first case of its type in the world, with global ramifications. The illegal downloading of movies, TV programs and music on a large scale was not disputed…

While many revel in the status quo – an illegal but uncontrollable free-for-all – the ruling sent a chill through the film, television and music industries, and indeed all holders of downloadable intellectual property. Copyright holders can and have sued individual users, but illegal file sharing is so pervasive that they could go broke trying to plug all the holes.

A High Court appeal is likely because of the impact of the case. It goes to the heart of global movie, TV and music distribution, which in its prime was a system of astonishingly profitable cartels. However, ISPs’ viability would be at risk, too, if they were to be liable for all independent user activities. The court ruling confirms that the system for protecting copyright and intellectual property is broken; the law has failed to keep pace with the internet.

Content providers are entitled to look to governments for a legislative answer, although great care must be taken not to trample online freedom. Britain and France are among countries considering a ”three strikes” rule to force ISPs to warn serial offenders against illegal downloads, with eventual termination of their accounts. The problem is that internet access is seen as a basic right in modern society.

Still, if illegal downloading of ”free” content continues unabated, everyone will suffer the consequences. Generating high-quality content involves costs, which are often considerable, and if creators of content cannot make money the flow will cease. This applies whether the content is film, television, music or, indeed, journalism. The twin legislative and market failures challenge content providers and users alike. Greed, whether for profits or ”free” material, may kill the goose that lays the golden egg.

Both the law and business models must evolve to take account of the new means of distribution and user expectations, so material becomes legitimately available online. That may require lower prices offset by higher sales volumes. The music industry has begun to embrace a download model. There are no easy answers, and no law can provide them. The one certainty is that any business that clings to the past and resists the new realities of an online world will be history.”