Skip to Main Content
The Liability of Internet Intermediaries The Liability of Internet Intermediaries

Contents

The Liability of Internet Intermediaries The Liability of Internet Intermediaries
1

The scope of secondary liability for copyright infringement 6.04

1.1

The meaning of authorisation 6.07

1.2

Joint tortfeasorship 6.31

2

Application to internet intermediaries 6.48

2.1

Platforms 6.49

2.2

Hosts 6.64

2.3

ISPs 6.66

2.4

Gateways 6.77

2.5

Marketplaces 6.83

3

Graduated response obligations 6.85

3.1

Overview 6.88

3.2

Evaluation of benefits and costs 6.105

3.3

Proportionality 6.142

6.01 Copyright, it is said, is one of the great balancing acts of the law.1 The rights it confers embody the basic tension between encouraging optimal creation and consumption of works; their boundaries reflect delicate compromises between creators, consumers, disseminators, and many other interest groups. These conflicts are exemplified in the enforcement of copyright against internet intermediaries. Copyright owners assert that almost one-quarter of global internet traffic,2 80 per cent of YouTube videos,3 and 97 per cent of BitTorrent transmissions infringe their copyrights.4 Although the prevalence of unauthorised content appears to be declining with the growth of legitimate services,5 digital piracy remains widespread. The services responsible for routing, storing, and processing these data deny responsibility for policing infringements, citing the impracticability of monitoring and their inability to adjudicate claims of infringement, while internet users fear disproportionate interferences with privacy, internet access, and innovation.

6.02 This chapter analyses the secondary liability of internet intermediaries for copyright infringement. It begins by considering the two main forms of secondary liability in copyright: authorisation and joint tortfeasorship. It then analyses their application to the main categories of internet intermediaries. Like the concept of ‘publication’ in defamation actions, copyright is beginning to develop built-in limitations which shield network-layer intermediaries from monetary liability. However, statutory interpretation often provides limited guidance to courts because in many instances copyright legislation reflects a position characterised by deliberate ambiguity and inarticulate compromise.6 This chapter does not consider the scope of primary liability for copyright infringement, which is dealt with in other specialist works.7

6.03 Recognising that secondary monetary liability can be a blunt tool for enforcing copyright online, various schemes have been enacted or agreed which provide for action by ISPs against their subscribers. Section 3 considers the solution embodied in the Digital Economy Act 2010, which requires ISPs to record and forward allegations of infringement, and contemplates technical measures to disconnect or suspend the accounts of repeat infringers. The relatively recent remedy of blocking access to copyright-infringing internet locations is considered in chapters 14 and 15.

6.04  Source of liability. Copyright is now a creature of statute.8 As such, the scope of property rights enjoyed by copyright owners—and the limits of others’ liability for their infringement—fall to be determined foremost by legislation. Section 16(2) of the Copyright, Designs and Patents Act 1988 provides that:

Copyright in a work is infringed by a person who without the licence of the copyright owner does, or authorises another to do, any of the acts restricted by the copyright.9

6.05 The copyright owner (or its exclusive licensee) has a statutory cause of action against anyone who infringes its copyright.10 Relevantly, the acts restricted by copyright include copying a work—for example, by storing it ‘in any medium by electronic means’11—and communicating it to the public by electronic transmission.12 The meaning of these acts in digital environments has been the subject of extensive judicial13 and academic14 treatment. This chapter does not examine the liability of parties who engage in restricted acts by reason of their own conduct;15 nor does it examine so-called ‘secondary infringements’ which arise from certain dealings with infringing articles16—both being examples of primary liability.

6.06 Instead, this section considers two doctrines which allow liability to be attributed for secondary wrongdoing: authorisation and joint tortfeasorship. Their effect is to augment the boundaries of responsibility beyond the defendant’s own exercise of the acts restricted by copyright to encompass acts committed by third parties. The following sections consider the scope of these doctrines and their application to internet intermediaries.

6.07  Origins. The clearest mechanism by which one person may become liable for the infringing acts of another is when that person ‘authorises another to do’ the relevant acts. This formulation dates to the Copyright Act 1911 (Imp),17 which defined copyright to include the exclusive right to engage in various restricted activities ‘and to authorise any such acts as aforesaid’.18 Once enacted, authorisation liability became a ‘separate species’ of infringement that existed in parallel with theories of direct infringement, joint tortfeasorship, and vicarious liability.19 It is one of ‘a range of concepts and devices’ from the 1911 Act to survive in modern legislation.20

6.08  Evolution. Authorisation liability has undergone a long and ‘tortuous’ evolution.21 Like the concept of publication, its meaning has been progressively enlarged in response to new technologies for reproducing and disseminating information—in particular, creative and industrial works—followed by periods of contraction and rebalancing. In copyright cases, that evolution can be divided into three main phases:

(a)

during the first phase, authorisation liability was almost non-existent, with courts recognising only a form of limited vicarious liability that required a relationship of agency with the primary infringer;

(b)

following the 1911 amendments, courts focused on control over the place of infringement by retail intermediaries; and

(c)

more recently, as manufacturers armed domestic infringers with reproduction technologies, concepts of authorisation began to emphasise control over the means of infringement.

Relatively few cases have considered the liability of internet intermediaries. This makes it important to consider this evolution in some detail.

6.09  Causation. Until the 1911 Act, statutory liability attached to those who engaged in the infringing activity or ‘caused’ it to occur.22 In practice, only its most direct causes were liable. Thus, in Russell v Briant, Wilde CJ described liability as confined to one who ‘by himself or his agent...actually takes part in a representation which is a violation of copyright.’23 This was essentially a limited doctrine of vicarious infringement under which a defendant could be liable only for the acts of his servants and agents. Simply to supply ‘some of the means’ of infringement—such as premises or equipment—was insufficient.

6.10 To similar effect, Karno v Pathé Frères refused to extend liability to sellers of cinematograph films who knew they would be tortiously exhibited by third parties. The trial judge commented that the claim ‘ought to have been brought, not against the defendants, but against the actual proprietors of the piratical performance impugned’.24 Such a narrow approach to causation reflected a view of infringement in which secondary liability had little room to operate.

6.11 By abandoning ‘cause’ for ‘authorise’, the 1911 Act widened the ambit of secondary liability. Its object was plainly ‘to sweep away’ decisions such as Karno.25 Thus, one writer argued that

the new Act has enlarged the protection accorded to the owner of a copyright by thus making it part of his monopoly to ‘authorise’ any of the acts referred to . . .26

6.12  Initial reception. Upon the enactment of the 1911 Act, academic opinion was divided27 and early cases interpreted authorisation liability narrowly. In Performing Right Society Ltd v Ciryl Theatrical Syndicate Ltd, Scrutton LJ considered that the words inserting authorisation ‘are superfluous and add nothing to the definition’ of copyright,28 a view repeatedly endorsed.29

6.13  ‘Grant or purport to grant’. Just two years later, an identically constituted Court of Appeal held, on facts virtually indistinguishable from Karno, that the defendant film vendor was liable for authorising infringing screenings by a third party cinema.30 Only Atkin LJ directly addressed the meaning of authorisation; Bankes and Scrutton LJJ concluded that the trial judge’s findings were justified on other grounds.31 In obiter, Bankes LJ observed that resort might be had to the ‘ordinary dictionary sense of “sanction, approve, and countenance”’;32 Atkin LJ preferred the narrower sense of ‘to grant or purport to grant...the right to do the act complained of’.33 Later courts are therefore correct to prefer the view of Atkin LJ, regarding Falcon as decided on the basis that one who hires out a film to a cinema ‘plainly’ purports to grant the right to screen the film.34

6.14  Policy considerations. One policy reason for the courts’ willingness to impose liability onto proprietors and vendors was the relative difficulty of bringing suit against the nomadic and impecunious parties who exhibited infringing performances: ‘A band’, it was said, ‘is often a migratory thing, and an action against it only might be of small avail to the plaintiffs.’35 In other words, where it was difficult to identify and proceed against primary wrongdoers, an intermediary may stand in a better position to avoid—and, if necessary, compensate—the copyright owner’s loss.

6.15  Nature of liability. However, the criterion described in Falcon proved difficult to apply,36 and later decisions oscillated between the available definitions.37 Despite disagreement about its scope, authorisation liability came to be accepted as derivative in nature and conceptually distinct from the primary infringer’s liability.38 This follows from the statutory language—which requires the ‘do[ing]’ of one of the restricted acts—and from the fact that if the primary actor’s conduct is not tortious or never actually takes place, there can be no actionable wrong in authorising it.39

6.16 A trio of cases in the early 1980s rejected attempts to hold manufacturers and sellers of cassette recording equipment liable for acts of copying by purchasers.

6.17 In the first case,40 a claim against a manufacturer of blank cassette tapes was struck out as disclosing no cause of action. The defendant had commissioned and published provocative advertisements in which various infringing uses of its cassette tapes were depicted. Despite those invitations, Foster J held that there was no evidence that the manufacturer had authorised any specific act of infringement by retail consumers who purchased the tapes.41

6.18 The mischief concerned by A&M Records was ‘home taping’ by consumers who used tapes to make infringing copies of musical works. In many ways, this practice closely resembles the modern phenomenon of internet file-sharing: both technologies offer individuals a cheaper alternative with an ‘extremely slender’ chance of detection by the copyright owner;42 in both instances, copyright owners sought to deploy authorisation liability to regulate those who supplied technologies that enabled copying; and in both cases, initial hubris gave way to sustainable revenue streams for right-holders in the form of lawful alternatives and new media.

6.19 The second case targeted the operator of retail record stores from which members of the public could borrow audio recordings and purchase discounted blank tapes in return for a subscription fee. Whitford J held that the defendant was not liable for infringing copies members made from loaned records because it had not purported to grant any right to engage in unauthorised home taping.43

6.20 One important aspect of Ames is that the Court cast aside the broad ‘countenance, sanction, and approve’ definition of authorisation, preferring the narrower view of Atkin LJ in Falcon. The defendant had only supplied source material and not equipment or premises;44 further, although it knew that borrowers may copy the cassettes, it did not grant permission to do so, instead warning borrowers to the contrary. In these circumstances, no inference of authorisation could be drawn.

6.21 In the third case,45 the House of Lords held that neither the manufacturer nor retailer of dual-deck tape recorders had authorised infringements by purchasers who used them to duplicate commercial recordings without permission. The speech of Lord Templeman in this case provides ‘[t]he fullest modern discussion’ of both the principles of tortious secondary liability and statutory authorisation liability.46

6.22  The facts. Again, the defendants had issued various ‘hypocritical and disingenuous’ advertisements extolling the virtues of their devices,47 which could duplicate tapes at twice the ordinary speed. Although usually accompanied by a small copyright disclaimer, they were intended to encourage purchasers to copy copyright-protected cassettes,48 and the defendants knew the devices would ‘inevitably’ be used for that purpose.49 Despite this, both the Court of Appeal and House of Lords rejected authorisation liability and endorsed Atkin LJ’s narrow test of authorisation.50

6.23  Lack of control over manner of use. Lord Templeman placed particular emphasis on the manufacturer’s lack of control over how its devices would be used, comparing Amstrad to the lending library in Ames. Simply to condone and even deliberately encourage an infringing act was not to authorise it, unless accompanied by conduct which suggested some authority to grant the permission required for infringing use. Consistently with Falcon, lenders and sellers would not authorise copying because they lacked control over primary wrongdoers and did not purport to grant any permission to do the acts complained of.

6.24  Rejection of liability in negligence. Notably, these cases rejected attempts to pin liability on intermediaries using the tort of negligence.51 Instead, they repeatedly affirmed the general principle that a person is not ordinarily under a duty to control another to prevent him from causing economic loss to a third party (here depreciation to the value of a chose in action). Although one person can clearly be liable in negligence for damage deliberately inflicted by another,52 such duties arise in ‘limited circumstances’.53

6.25  Rejection of duty to prevent infringement. Similarly, the courts refused to recognise a general duty not to infringe—which would be an illegitimate parallel duty to the statutory torts54—or not to facilitate a customer’s infringement—which would be ‘unsatisfactory’ and ‘far too heavy a burden’.55

6.26 Such a conclusion seems beyond doubt: to hold otherwise would extend liability in negligence for pure economic loss far beyond established principles56 and fracture the coherence of doctrines of joint tortfeasorship. In refusing to recognise an action for negligent facilitation, these decisions made clear that (outside limited cases of joint tortfeasorship) tort law will not assist a copyright owner to whom Parliament has not given a remedy.57

6.27 The act of authorisation need not take place within the United Kingdom.58 This arises from the fact that section 16(2) of the 1988 Act is not territorially restricted: it creates liability wherever a person authorises another to do any act restricted by the copyright. Although the primary acts of infringement must, to fall within section 16(1), be done ‘in the United Kingdom’, the same restriction does not apply to the place where the grant or purported grant of authorisation occurs.

6.28  Authorisation from abroad. In Records Inc v Music Collection International Ltd it was open to the claimant to argue that a Danish company had authorised from Denmark by licensing the right to make and sell certain infringing sound recordings in the United Kingdom.59 The defendant argued that, in the absence of express words or necessary implication, domestic statutes are presumed to be territorial and will not apply extraterritorially to persons who are not present in the United Kingdom.60 Copyright being territorial, the defendant argued that there was no reason to extend section 16(2) to acts done abroad such as the formation of a licence agreement.

6.29 The Court rejected this argument. As Hoffmann LJ explained, a sufficient jurisdictional nexus was created by the requirement that the act of primary infringement be done within the United Kingdom. This ‘overlap’ between the statutory torts of authorisation and primary infringement justified exercising exorbitant jurisdiction over preliminary acts performed abroad, because the consequences of the authoriser’s conduct were domestic infringements.61 To hold otherwise would lead to anomalies by permitting any licensors to execute documents abroad in order to avoid English authorisation liability.

6.30 It was common ground that this approach was correct in Football Dataco Ltd v Sportradar GmbH,62 where Austrian and German companies were alleged to authorise acts of reproduction and infringement of database right arising from the display of data on English users’ computer screens. This broad approach could apply to almost any infringement which occurs when material is viewed on a computer or other electronic device in the United Kingdom. This makes a plea of authorisation a potentially effective device to extend copyright liability to foreign internet intermediaries.

6.31 Although copyright infringement sounds in a statutory remedy, it is a tort63 to which common law principles of joint tortfeasorship apply just as they do to all other torts.64 The two recognised connecting factors are procurement and common design. These are not independent torts, but instead create joint liability for the primary infringement.65 Joint liability might also arise based upon principles of agency and vicarious liability, though the principles are unsettled.66

6.32  Relationship to authorisation liability. Authorisation often turns on identical factors to joint tortfeasorship, which suggests significant overlap and a degree of redundancy. Indeed, the difference may only be significant where a service provider is established abroad and has formed, but not undertaken, any act in ‘furtherance’ of a common design with a local primary infringer; in such a case, authorisation may be arguable but probably not joint tortfeasorship from abroad of a tort in England.67 Where joint tortfeasorship is alleged in relation to foreign acts of primary infringement that infringe foreign rights, the conventional view appears to be that the law of that country will apply to determine whether secondary liability is established.68

6.33  The test. As is explained in chapter 5, it is ‘well established’ that A, a service provider, is liable as a joint tortfeasor for procuring copyright infringements carried out by B.69 The modern test is expressed in CBS, where Lord Templeman explained that procurement is both causative (A’s ‘inducement, incitement or persuasion’ must cause B’s infringing act) and intentional (A must intend or ratify B’s infringing act).70 This bears some resemblance to inducement theories of secondary infringement recognised in the United States.71

6.34 On the facts of CBS, procurement failed because purchasers infringed of their own accord, not because they were persuaded by advertisements; Amstrad’s conduct therefore made no causal contribution to the infringements.72

6.35  Subject matter of procurement. In CBS, the Court held that the act of procurement must not be ‘at large’.73 If this is interpreted as meaning that the secondary party must procure a specific infringement, then it would set a high bar for claimants who wish to pursue the providers of technologies used for large-scale infringement. Recognising that to plead and prove each primary infringement is usually impracticable, as it was in CBS and A&M Records, modern cases tend to take a more pragmatic approach.

6.36  Incitement. Procurement is also said to require incitement, which requires the secondary wrongdoer to have ‘made himself a party to the infringement’.74 The authorities do not clarify precisely what this involves, but it appears to involve an assumption of responsibility for infringing uses of technology, and may be evidenced by a deliberate request or instruction that the technology should be used for infringing purposes.

6.37 This is a high threshold. Merely supplying the equipment, even with knowledge and intent that it be used tortiously, is insufficient if it is capable of both infringing and non-infringing uses. For such dual-use technologies, the only ‘party’ to the infringement is the individual who chooses to engage in the infringing uses. This is consistent with the view that, copyrights being monopolies in restraint of trade, they should not be given de facto enlargement by holding third parties liable for ‘acts short of infringement’.75

6.38 The second connecting factor is participation in a common design. As is discussed in chapter 5, this refers to agreed or concerted action to carry out tortious acts (here infringing the copyright). Several issues arise in relation to internet intermediaries.

6.39 For a common design to be established, the defendant must agree to participate in the infringing scheme and, although the agreement may be tacit,76 it cannot arise by indifference or omission.

6.40  Substantiality. The involvement of the secondary party in the primary infringement of copyright must not be trivial or de minimis. Initially, the threshold of participation was high: in Lyon v Knowles a share in the proceeds of an infringing production was insufficient for a theatre proprietor to be jointly liable with its lessee without capital contributions or ‘joint action or control over the performances’.77 Now sufficient involvement may be inferred from the existence of an economic interest in the infringement, and even that is unnecessary.78

6.41  Scope of agreement. Because the agreement to engage in tortious activity need not be express, the existence of a contract between a service provider and third party infringer has been said to be ‘irrelevant’ unless its terms contemplate infringement.79

6.42 In CBS, Lord Templeman concluded that the defendant did not act in pursuit of a common design partly because its recording devices were dual-use machines. Purchasers’ decisions to engage in the infringing uses were their own and not referable to any agreement.80 This reflects the principle that simply supplying a lawful ‘means’ of infringement is insufficient. For similar reasons, an independent IT consultant is not liable for supplying and installing computer hardware subsequently used to upload infringing material.81

6.43  Services without lawful uses. Conversely, supplying a service which has no lawful uses may suggest a common design to engage in an unlawful use.82 In Douglas, a magazine editor commissioned photographs of the claimants in circumstances where the photographer would necessarily have to use unlawful means to fulfil the commission.83 This was enough to make the magazine and editor engaged in a common design with the photographer to interfere with the claimants’ privacy. This arrangement was unlike the ‘lawful provision of lawful facilities’ in CBS.

6.44  Apportionment. To prevent the injustice that might otherwise arise from imposing joint liability which is disproportionate to a secondary party’s participation, secondary infringement doctrines operate to prevent service providers from facing liability for contributory conduct which is insufficient to justify imposing full responsibility for the acts of others. In this sense, such doctrines are normative gatekeepers.

6.45 Once defendants are admitted within the gateway of liability, responsibility may be apportioned by means of an order for contribution as between the primary and secondary infringer. However, there is no guarantee of proportionality. This is especially problematic where the other tortfeasors are anonymous or likely to be judgment-proof. In these circumstances, it might be unfair to hold one service provider liable for the full extent of the loss caused by a third party’s infringement.

6.46  Absence of a practical remedy. If an intermediary’s conduct falls short of the threshold required for joint or primary liability, then the claimant may be left without any remedy at all. When Amstrad was before the Court of Appeal, Nicholls LJ expressed his ‘profound dissatisfaction’ at dismissing the claimants’ appeal. Although the appeal was ‘misconceived’, the problem remained of intermediaries who

are, on a large scale, inciting others to infringe copyright in circumstances where the copyright owners have no practical remedy against the actual infringers, and there is nothing the copyright owners can do through the courts to stop them. If, indeed, that is so, the present state of the law is, in my view, gravely defective.84

6.47 The same criticism can be made with even greater force of the widespread infringements occurring online today. In practical terms, there is very little copyright owners can do to stop them, at least using the machinery of legal remedies: as the High Court commented in iiNet, suing individual infringers would be ‘somewhat impractical’85—a mere ‘teaspoon solution to an ocean problem’.86 While the answer is not necessarily to impose monetary liability on service providers, doctrines of secondary infringement arguably fail to supply a sufficiently nuanced solution.

6.48 Doctrines of secondary copyright liability have regularly been invoked against application-layer and network-layer internet intermediaries. This section examines cases involving their most common targets—platforms, hosts, and ISPs—before speculating on the potential liability of gateways and marketplaces.

6.49 The cassette recording cases substantially narrowed the scope of authorisation liability faced by secondary parties. By insisting upon an express or implied grant of a right to engage in the infringing activity, these decisions made all but the most egregious forms of calculated facilitation insufficient. Whether the alleged authorisation consists of goods sold (as in Amstrad) or ongoing subscription services (as in Ames), all conduct short of purporting to grant permission cannot support an authorisation claim. Importantly, the decision in CBS did leave open the prospect that such a grant can be implied from the circumstances and conduct of the defendant—a possibility readily taken up in three later cases involving website operators.

6.50 As the leading authority, Newzbin87 suggests a broad, multi-factorial approach to authorisation by application-layer services.

6.51  Facts. The defendants operated Newzbin.com, which compiled ‘reports’ identifying and aggregating files hosted on third parties’ Usenet servers. Premium members (who paid a subscription fee) could find and download files in a single bundle known as an ‘NZB’ package, which obviated the need to download hundreds of smaller archives and formed the ‘crucial element’ of the service.88

6.52 The defendant claimed that its index was ‘content agnostic’, but the vast majority of reports (greater than 99.7 per cent) related to infringing cinematograph films, software, and other copyright works. Although Newzbin had copyright policies and warnings, it did not filter protected works; warnings were ‘entirely cosmetic’ and mere ‘window dressing’.89 Its operators knew that infringement was widespread. However, none of the infringing material was actually hosted by Newzbin and it could not directly control which files its members chose to access.

6.53  Test of authorisation. Despite purporting to adopt the narrow grant-based test of authorisation, Kitchin J applied a broad multi-factor test of implied authorisation, and regarded the Australian authorities (which, until iiNet, preferred Bankes LJ’s synonym test) as ‘entirely consistent’ with English principles.90 While ‘mere enablement, assistance or even encouragement’ was said to be insufficient, Newzbin had authorised infringement having regard to five factors:

(a)

Contractual relationship. First, although not decisive, the operator of the Newzbin website had an ongoing contractual relationship with primary infringers.

(b)

Means of infringement. Second, the website created and supplied the ‘means of infringement’, a sophisticated NZB browsing interface whose structure and features were directed at locating and copying copyright works.

(c)

Inevitability of infringement. Third, when used as intended, the NZB facility ‘inevitably’ infringed copyright.

(d)

Control over use. Fourth, Newzbin maintained control over its use, which could be terminated at any time.

(e)

Failure to take reasonable steps. Finally, despite knowing that most content infringed copyright, it did not take any steps to filter or discourage reports of copyright material.

6.54 In these circumstances, Kitchin J inferred

that a reasonable member would deduce from the defendant’s activities that it purports to possess the authority to grant any required permission to copy any film that a member may choose from the Movies category on Newzbin . . .91

6.55  Criticisms. Although liability may be the appropriate result on these facts, to impose it on the basis that members would wrongly assume a permission to copy where none in fact existed has the appearance of fiction. Given the low cost of membership and the website’s many references to copied material, it seems doubtful whether anyone but the most naive members would assume Newzbin possessed any more authority to grant permission than they did themselves. Certainly, this is consistent with evidence of members’ concern at being personally identified: they knew all too well that what they were doing was unauthorised by the relevant copyright owners. Moreover, it may lead to absurd results: if a website dealing exclusively in pirated content states prominently that it possesses no authority to grant a licence to copy the works referenced there, the operator should still clearly be liable.

6.56  Joint liability of platform operator. This scenario is better dealt with using joint tortfeasorship. In Newzbin, the website’s structure, contents, and design signalled a common design to infringe the claimants’ copyrights; the defendant also ‘induced’ its editors to include meticulous listings of copyright works, guided members to the location of particular works, and profited from subscription fees.92 Because the service inevitably caused infringing copies to be made, unlike the tape recorders in CBS, and Newzbin knew infringement was widespread, the inescapable inference was that Newzbin made itself party to its members’ infringing acts.93

6.57  Evidence of specific infringements. Although the claimants could not adduce evidence of specific infringements by particular members, this was only because the defendant retained no records of transmissions. Kitchin J made clear that this was not fatal to a finding of joint tortfeasorship but merely one factor to be considered.94 This reflects a pragmatic concession to the traditional approach.

6.58 In this case,95 the liberal view of authorisation recognised in Newzbin was applied to hold the operators of a notorious BitTorrent tracker, The Pirate Bay (‘TPB’), liable for authorising infringements by their United Kingdom users.

6.59  Relevant factors. Since Dramatico was also a case concerning authorisation by supply of hyperlinks to infringing materials, Arnold J referred to the same five factors as Kitchen J in Newzbin. Again, TPB supplied a ‘comprehensive service’ which allowed its users to search for and download torrent files using BitTorrent. The Court characterised the torrent files as the means of infringement, since—analogously to NZB files—they allowed the many fragments of a copyright work to be reassembled. The defendant supplied this means to users of the TPB website.96

6.60 Further, infringement was both inevitable and intended by TPB, as suggested by its name, mission statement, and provocative media comments. The existence of non-infringing material was assumed to be trivial. As to control, TPB could prevent infringement by removing torrent files, which it did on occasion for other harmful materials, but chose not to do ‘as a matter of policy’ for copyright materials. The operators took no steps to prevent infringement, instead assisting users to circumvent blocks instigated by their ISPs. Arnold J held that these facts ‘go far beyond’ mere assistance, and satisfied both the broad and narrow tests of authorisation.97 The case was, as Arnold J noted, an even stronger one than Newzbin.

6.61 Again, however, it seems difficult to describe TPB’s service as a purported grant of authority to copy, though it clearly involves procurement and common design. Like Newzbin, TPB’s operators structured their website around infringement and made users into inevitable infringers. Common design therefore provides another basis for liability: the inescapable inference from their facilities having no (or de minimis) non-infringing uses was that they agreed their users could infringe. Arnold J accepted the ‘induce, incite or persuade’ test from Newzbin and held that TPB’s operators were also jointly liable as procurers.98

6.62 In this Australian decision,99 cited with approval in Newzbin, the operator of a website linking to MP3 files of the applicants’ musical works was held to authorise users’ infringements. The Court rejected an analogy with the manufacturers of blank tapes: Mr Cooper may not have had precise control over which files were accessed by users of his website, but he could choose not to make the ‘technical capacity’ for infringement available and to design the website differently.100 Respectfully, such a broad test of control makes the concept meaningless as a limiting criterion, since it would mean that any developer of a new technology has control in the loose sense that they could choose not to make the technology available.

6.63  Joint liability. Although Cooper presents another clear example of calculated facilitation, it is again more properly classified as joint, rather than authorisation, liability, though the Court concluded that neither service provider was a joint tortfeasor with end users for the puzzling reason that the index of hyperlinks was not edited or compiled by the defendants.101

6.64 No English cases have considered the position of those who host infringing materials. There are a number of reasons for this dearth of decided cases. First, the hosting safe harbour has immunised those who do no more than store infringing material.102 Second, notice-and-takedown procedures are now commonplace, leading many hosts to remove obviously infringing content when requested. Third, the overwhelming majority of infringing material is stored in overseas jurisdictions where storage is cheaper and hosts less scrupulous. Injunctive relief in England may be difficult to enforce abroad.

6.65 Where hosting goes beyond passive storage, a host may be liable for authorising infringement. In Cooper, the host (‘Comcen’) assisted Mr Cooper to create the website, offered free hosting in exchange for advertising, and thereby enjoyed a commercial benefit from the high volume of infringing traffic it knew the website received.103 The Court held Comcen liable as an authoriser because it was ‘complicit in’ the means of infringement,104 took ‘an active role’ and had control.105 This reasoning is subject to the criticism that the ‘means’ of infringement could be almost any antecedent technical condition, from computer hardware to electricity. It offers no principled criterion to limit the scope of authorisation liability. Further, like defamation cases, the distinction between an ‘active’ and ‘passive’ role remains unclear, but presumably requires more than a financial benefit, which almost all service providers receive.

6.66 No English court has directly considered the prima facie liability of ISPs for authorising infringement by their users. However, the Australian decision in Roadshow Films Pty Ltd v iiNet Ltd [No 3] provides helpful guidance.106

6.67  The facts in iiNet. In that case, iiNet, a national ISP in Australia, was held not to be liable for authorising its customers’ BitTorrent downloads of the applicants’ cinematograph films from third parties. The applicants argued that iiNet should be taken to authorise those infringements unless it took steps to warn or disconnect the relevant customers and remove infringing material from their computers.107 In effect, they put their case as breach by omission of a statutory duty of care—the kind of duty previously rejected in CBS.108 More than half of iiNet’s internet traffic consisted of such file-sharing. Although its customer agreement prohibited it, iiNet took no steps to disconnect or suspend customers whose accounts were implicated in notices sent by the applicants’ investigators.

6.68  Decision at trial. At trial, Cowdroy J dismissed the claim on various grounds, including that iiNet did not provide the actual means of infringement (BitTorrent)—over which it had no control—but merely an essential precondition (namely internet access). His Honour reread the Australian authorisation authorities as identifying the ‘true means’ of infringement: a party who provided a ‘necessary precondition’ to infringement would not necessarily be liable as an authoriser, whereas one who provided the actual ‘means’ could be.109 Although this reasoning was rejected on appeal as unsupported by authority, the High Court of Australia unanimously agreed that iiNet had not authorised its subscribers’ infringements.110

6.69  High Court decision. Like Cowdroy J, the majority judgment distinguished BitTorrent from other P2P distribution systems such as Grokster and KaZaA, twice observing that both BitTorrent and the internet had ‘diverse’ non-infringing uses111 and that iiNet’s services were not ‘bound’ to infringe.112 While the provision of access was a necessary precondition of infringement, to download copyright materials via BitTorrent involved several additional steps: account holders must first download a BitTorrent client, then find a ‘torrent’ file, connect to a tracker server such as TPB, locate peers, and download and reassemble file pieces. iiNet had no ‘direct technical power’ to prevent any of these activities, and did not supply any of these further means. Further, it could not remove infringing material once downloaded or filter its retransmission.113

6.70 The plurality judgment of Gummow and Hayne JJ reached the same conclusion. Relying on CBS, their Honours emphasised that the makers of copying technology owed no duty ‘to prevent or discourage or warn against infringement’ and authorisation could not be repurposed to achieve that end.114 Not only would such a duty be wide and uncertain, but it would be at iiNet’s expense and would grant copyright owners a remedy (disconnection) which they could not have obtained in actions against primary infringers.

6.71  Lack of control over subscribers. Both judgments emphasised iiNet’s minimal control over subscribers; while it could exercise ‘attenuated’ control by activating its contractual power to disconnect, it was not unreasonable to refuse to do so on the basis of incomplete allegations, or to refuse to monitor subscribers’ activities.115

6.72  Reasonableness of disconnection. The majority accepted that evidence falling short of civil proof of infringement will not supply a reasonable basis for sending warning notices or disconnecting customers. To require an ISP to act on less ‘assumes obligations...which the Copyright Act does not impose’.116 Implicit in this reasoning is a judgment that ISPs should not be required to disconnect customers or undertake further investigation into allegations of infringement at their own expense.

6.73  Test of authorisation. The Court unanimously adopted Atkin LJ’s formulation of authorisation, suggesting that Australian case law is reconverging on a standard approximating CBS. The narrower formulation in Falcon emphasises ‘the immediacy...of the relationship between the primary infringement and the secondary infringement’.117 In recasting authorisation liability in narrower relational terms, their Honours departed from use of synonyms, which illegitimately ‘expand the core notion of “authorise”’ by seizing upon a broader approximation of the word.118 iiNet’s supply of internet access did not involve any grant of a right to use the internet for tortious purposes; to the contrary, the customer agreement ‘indicated iiNet’s express, formal and positive disapproval’ of such use.119 Without ‘direct power’ to prevent the primary infringements, merely countenancing—even encouraging—them was insufficient.120

6.74  Applicability in the United Kingdom. The iiNet first instance decision was cited with approval in Newzbin.121 Despite obvious differences between the English and Australian tests of authorisation—in particular, section 101(1A) of the Australian Copyright ActiiNet is a persuasive authority, particularly because questions of authorisation liability are not harmonised under the Information Society Directive. If followed in England, it seems likely that, even when confronted with specific knowledge that a subscriber is using the internet to infringe copyright, an ISP will not bear authorisation liability for infringements committed by its subscribers using tools supplied and controlled by third parties. However, it may come under a duty to block access to the infringing website—a possibility discussed in chapter 14.

6.75  Safe harbours. In any case, the mere conduit safe harbour is likely to provide complete protection from monetary liability arising from transmission. This question is the subject of a pending reference before the Court of Justice in McFadden, which is discussed further in chapter 12.122

6.76  Criticism. Some commentators have criticised the de facto immunity of ISPs from subscriber copyright infringement as unjustifiably removing incentives to discourage infringement or adopt policies for dealing with persistent infringers.123 However, the backstop of tortious secondary liability should not be forgotten: tacit agreement or intentional inducement could still create liability. Further, the duty to block access to infringing materials in many cases provides a compelling alternative to monetary liability. Alternatives to monetary liability, such as notification regimes, are discussed in section 3, while blocking injunctions are discussed in chapters 14 and 15.

6.77 Even more limited guidance is available about the secondary liability of search engines and other gateways. Two general observations may be made.

6.78  General purpose gateways. First, it seems clear that a search engine which does no more than index generalist content, without emphasising specific infringing materials, will not authorise infringement by those who happen to access infringing content. Simply to direct users to available third party content is not to ‘purport to grant...a right or licence’ to copy it (other than perhaps by making temporary copies).

6.79 Universal search engines, which facilitate access to third parties’ materials in a more or less neutral manner, stand much closer to the position of a device manufacturer or lending library than to someone who specifically permits infringing conduct: indeed, the search engine is generally not the tool with which the infringement takes place—merely a precondition of infringement—and not even an essential means (since the infringing materials could be accessed directly or via another search engine). To the extent infringing materials are themselves displayed in search engine snippets, the engine is arguably an automaton engaging in ‘involuntary’, non-tortious copying.124

6.80  Gateways dedicated to piracy. Second, a search engine which is designed to index infringing material can be liable if its knowledge and technical control, coupled with indifference or inaction, ‘reach a high degree from which authorization or permission may be inferred’.125 Applying Ames and CBS, mere encouragement (eg automatically generated search suggestions) or benefit may not be enough, but whether authorisation can be implied would ultimately be a question of fact.126

6.81  Mere hyperlinking. Some authorities suggest it is possible to form a common design by hyperlinking to a website which provides a service that ‘inevitably’ results in infringement. In Football Dataco Ltd v Sportradar GmbH,127 a Gibraltarian website operator linked to Sportradar’s live football scores service, which incorporated data sourced from the claimants’ databases. The website used the scores service to attract punters to its own sports betting website; it was presented as the default option and in such a way that infringement of the database rights was inevitable by anyone who accessed it.

6.82 Referring to this encouragement, Floyd J held that the website operator was not a ‘mere intermediary, like eBay’; its activities went beyond facilitation and ‘adopt[ed]’ the acts of data extraction in pursuit of a common design with website users.128 By contrast, Sportradar merely made the data available in Austria to website operators, which was insufficient to form a common design with English website users who accessed the data. Importantly, no party knew that its use of the service was tortious, which reflects the principle that it is sufficient for an agreement to relate to activities rather than the specific commission of wrongdoing. However, outside special cases it seems doubtful that hyperlinking alone would be sufficient inducement or agreement for joint liability.

6.83  Payment processors. American courts have dismissed claims against credit card gateways because they have ‘no direct connection’ to infringement.129 Simply making copying more probable or straightforward is an insufficiently material contribution for secondary liability; further, there is a strong public interest in insulating financial intermediaries from threats of liability for processing transactions.130

6.84 The same result seems likely under the English doctrine of authorisation, since payment processors and auction websites generally do not grant the right to copy or use purchased matter in any particular way simply by acting as the vendor’s agent. In particular, most marketplaces lack actual control over how their transaction systems are used, do not encourage or induce infringement, and take no steps to promote those systems as means of infringement. Certainly, this is consistent with Floyd J’s description of eBay in Sportradar. Again, however, it is possible to conceive of marketplaces which are dedicated to the sale of infringing goods, unauthorised parallel imports, or which take steps to intervene in and promote specific infringing sales. Such activity may eventually reach a degree from which it is sufficient to infer authorisation or a common design.

6.85  Policy context. Wielding a powerful rhetoric of artists as victims, piracy as theft, and intermediaries as free-riders,131 copyright owners have pressured regulators to supply more effective remedies to combat online infringement. These attempts to alter the ‘legal and procedural “rules of the game”’ have proven highly successful,132 culminating in a worldwide shift towards non-monetary enforcement schemes.

6.86  Enforcement schemes. These take many forms, ranging from private notice-and-notice agreements to statutory ‘three strikes’ regimes such as the Digital Economy Act 2010 (UK) (‘2010 Act’). Their common element is that ISPs are required to take some action against subscribers accused of infringing copyright.133 Underlying concerns about their necessity, proportionality, and compatibility with human rights are two basic questions: the extent to which internet intermediaries should be compelled to internalise the costs of infringement; and the legitimacy of non-monetary penalties levied upon subscribers, such as disconnection.

6.87 This section outlines ISPs’ obligations under the scheme established (but not yet implemented) by the 2010 Act. It then assesses the scheme’s proportionality and compatibility with safe harbours and general limits on copyright remedies.

6.88  Origins of ‘graduated response’. The 2010 Act creates a ‘soft’ mechanism for digital copyright enforcement that requires ISPs, right-holders, and an industry regulator (Ofcom) to operate a scheme by which primary infringers can be identified, educated, and penalised. Graduated response originated as a theory of nuclear deterrence that recommended escalating nuclear payloads to defeat threats of aggression.134 This forgotten relic of the missile age was revived largely by sustained lobbying135 and the failure of ISPs and copyright owners to agree on self-regulatory anti-piracy measures, and who pays for such measures.136

6.89  Main components of the scheme. The scheme comprises three main components:

(a)

first, the framework of regulatory powers and duties contained in the 2010 Act itself;

(b)

second, the Initial Obligations Code (‘Code’),137 in which Ofcom specifies details of the scheme’s operation;138 and

(c)

third, ancillary statutory instruments, including a cost-sharing order and other instruments issued by the Secretary of State.139

6.90  Nature of obligations. While these instruments create obligations which are not ‘duties’ in the sense familiar to private lawyers and do not sound in ‘liabilities’ when breached, they represent a new class of ‘soft’ administrative obligations owed to regulators. In this sense, the 2010 Act reflects a deliberate political compromise between the interest structures of copyright owners, ISPs, and consumers. However, as a deliberate compromise, it leaves crucial features of the scheme undetermined and fails to clarify its relationship with existing remedies.

6.91  Qualifying ISPs. The scheme applies only to ‘qualifying ISPs’—those with more than 400,000 fixed-line subscribers who obtain internet access ‘under an agreement’ and using an allocated IP address.140 Mobile and wholesale ISPs are excluded—the former on the assumption that infringement is less prevalent141 and more difficult to trace to individual subscribers;142 the latter to ensure that hosts, backbone operators, and application-layer platforms are not captured. Currently, the only United Kingdom ISPs who exceed the qualification threshold are BT, Sky, Virgin Media, TalkTalk, and EE.

6.92 The clear intention is that the scheme should apply only to large retail ISPs, sparing smaller businesses disproportionate costs of compliance. Although providers of wireless access—for example, in coffee shops, libraries and universities—will fall within the definition where service is provided under an agreement (and not gratuitously), most wireless ISPs would be unlikely to meet the quantitative threshold. Nevertheless, in 2010 qualifying ISPs represented at least 93.5 per cent of the United Kingdom residential market for internet access.143 Their combined market share has since fallen to around 90 per cent.144

6.93  Qualifying copyright owners. Only a ‘qualifying copyright owner’ may avail themselves of the scheme. These are copyright owners who have given advance warning of their intention to issue notices to ISPs and paid any required share of the anticipated enforcement costs.145 Although any copyright owner (or its agent) may participate in the scheme,146 in practice these requirements serve to confine its operation to repeat-player claimants in the music and film industries, further limiting the volume of complaints faced by qualifying ISPs.

6.94 The 2010 Act divides ISPs’ duties into two categories: the first, initial obligations arise upon the Code coming into force; the second, more onerous technical obligations start only once declared by the Secretary of State at least one year later. The initial obligations further divide into two duties.

6.95 First, qualifying ISPs must send notifications to their subscribers whenever a valid copyright infringement report (‘report’) is received from a qualifying copyright owner in relation to a subscriber’s IP address.147 Reports are sent electronically in an agreed format, and must include various particulars of the work and its alleged infringement.148 Unlike the notification standard in iiNet, copyright owners need not show proof to a civil standard—merely a belief that an infringement is occurring by means of the ISP’s service.

6.96 Upon receiving a valid report,149 an ISP must contact the person to whom it relates with a standard form notice.150 The form of notices is initially polite and educative, but becomes progressively sterner with each notification.151 Their ‘central purpose’ is educational: to inform subscribers that their conduct is tortious and persuade them to change their behaviour.152

6.97 Relevant ISPs must maintain a list of subscribers who are repeatedly accused of infringement. Any subscriber who is sent three successive notifications within a 12 month period must be added to a copyright infringement list. This list records the details and number of reports made against each subscriber, without revealing their identity.153 An ISP must also delete a subscriber upon the expiry of his or her earliest notification.154

6.98  Disclosure of anonymised lists. Relevant ISPs must supply the anonymised subscriber infringement list upon request to any qualifying copyright owner.155 This will enable the copyright owner to determine which subscribers have received the highest numbers of notifications within the relevant period.

6.99  Disclosure of de-anonymised lists. A relevant ISP may also be obliged by court order to disclose an unredacted copy of the repeat infringer list to a qualifying copyright owner. The 2010 Act does not prescribe any details of the preconditions or procedure applicable to an application for such disclosure. Although the intention appears to be to preserve Norwich Pharmacal disclosure, its relationship with the 2010 Act remains unclear.156

6.100  Parallels with the UDRP. The notification scheme established by the 2010 Act shares some parallels with the UDRP, which aims to provide a cost-effective procedure to determine allegations of abusive domain name registration (eg cybersquatting). Like allegations of P2P file-sharing, such allegations are overwhelmingly well founded.157 Further, like the 2010 Act, UDRP decisions are enforced by intermediaries (registrars) and unsuccessful parties may appeal. Conversely, determinations under the Code do not rest on jurisdiction by consent and are preliminaries to an actual claim determining the merits.

6.101 The government has indicated that it expects the initial obligations to reduce online infringement by 70 to 80 per cent,158 a target widely considered optimistic. If, as seems likely, the target is not met after 12 months, the Secretary of State can activate a second set of ‘technical’ obligations,159 which require ISPs to adopt various measures to deter and prevent infringement by repeat infringers. To avoid the risk of obsolescence, the obligations themselves are left to a future technical obligations code.160

6.102  Available technical measures. Measures could include: per-subscriber access restrictions to specified content or protocols;161 capping and traffic management, which limit the volume and speed of data, both infringing and non-infringing, which subscribers may access;162 and suspension or termination of a subscriber’s access.163 Little is known about the technical obligations at this stage. It is also unclear whether all qualifying ISPs would have the technical capability to implement each of these technical obligations without changes to their infrastructure.

6.103  Powers of Ofcom. ISPs’ obligations under the 2010 Act are enforced by Ofcom under a system of discretionary powers. It may direct ISPs and copyright owners to take specified remedial steps, impose civil penalties of up to £250,000 per contravention, and require payment of compensation to affected third parties.164

6.104  Nature of liability. Unlike private remedies, these penalties are neither compensatory nor restitutionary in nature, though they may include compensation. They are determined at first instance by administrative rather than judicial decision. Enforcement actions are brought by a public authority, Ofcom, rather than private claimants, and only after a process of investigation, notification, and failure to take remedial action.165 The costs of administering and implementing the scheme are discussed in chapter 18.166

6.105 As a system designed to reduce claimants’ enforcement costs, the 2010 Act falls to be assessed on a cost–benefit basis. This assessment is relevant both to the proportionality of the scheme as a whole, and to the proportionality of individual applications made by qualifying copyright owners under the scheme. This section identifies several potential benefits, but their magnitude is uncertain compared to the potential overall costs of compliance. Caution is therefore warranted until the impact of the initial obligations on ISPs and subscribers can be properly assessed.

6.106 Supporters of the scheme claim it will deter infringement, simplify enforcement, provide certainty to ISPs, and deliver a fair and impartial system for determining copyright complaints. The government estimates total gains at £1.2bn over ten years.167 However, upon closer examination many of these benefits appear doubtful or rest on untested assumptions.

6.107  Effectiveness. Three main benefits to copyright owners may be identified. First, notification provides a simpler and more targeted method of reducing internet infringement than legal actions against primary or secondary infringers.

6.108Ineffectiveness of traditional enforcement. A decade of end-user litigation has demonstrated its inability to deter insolvent, anonymous, and profligate defendants.168 Even clear verdicts against services such as Newzbin and TPB can be illusory: replacements emerge, hydra-like, while the unsuccessful defendant has meanwhile vanished into the mists of the internet, leaving the claimant to bear its costs and, like Sisyphus, begin enforcement anew.

6.109Considerations of fairness. If notification is effective, claimants would no longer need to sue their customers indiscriminately—a practice which has made recording labels ‘the most hated industry since the tobacco industry’.169 As Parker J commented in R (British Telecommunications plc) v Secretary of State, infringer lists are a ‘more efficient, focussed and fair system’, since they allow claimants to decide which primary infringers are worth pursuing.170

6.110Anticipatory enforcement. By focusing on the most active infringers, claimants can maximise the ratio of expected remedies to enforcement costs. However, the scheme ultimately relies on threats of sanctions against individual subscribers to be effective. If subscribers realise proceedings will not be instituted even against egregious infringers, the prophylactic effect of notices will quickly evaporate. Similarly, while receiving notices might function as a ‘symbolic reminder’ of copyright norms,171 it might also have the opposite effect if notices could be ignored without sanction.

6.111  Deterrence. Second, there is some evidence that sending notices to subscribers would materially reduce infringement levels. However, the available empirical evidence is divided.

6.112Studies suggesting substantial deterrent effect. One English survey reported that 70 per cent of respondents would stop infringing after receiving one notice; another 16 per cent upon receiving a second.172 This is consistent with two claimant-commissioned studies suggesting that a majority of consumers would stop downloading films if warned not to do so—71 per cent in New Zealand,173 81 per cent in Germany.174

6.113Evidence suggesting partial deterrent effect. Another English survey suggests only 33 per cent would respond to warnings without any threat of sanctions.175 Empirical evidence suggests at least some positive impact: the French authorities claim a 26 per cent reduction in infringing P2P activity,176 while South Korea reports a 70 per cent reduction after each successive notice.177

6.114Theoretical analysis. The empirical findings suggesting some level of deterrent effect are supported by theory. As Yu observes, notification (backed by effective threats of sanctions) can deter infringement in a similar way to the accrual of demerit points for drunk driving or speeding.178 Such sanctions alter users’ calculus of wrongdoing to increase the expected value of penalties for infringement.

6.115Impossibility of full deterrence. Although the expected cost of infringing behaviour could not be raised to a point sufficient to deter all infringement or dissuade sophisticated users from evading detection, this is not the objective: the 2010 Act only needs to maintain reasonable compensation for copyright owners. The fact that recidivist infringers can evade the system is not fatal to its efficacy unless evasion becomes widespread. Accordingly, Parker J’s conclusion in BT that the scheme ‘may well have [a] positive effect’ seems at least plausible.179 Nevertheless, the extent of its effectiveness remains speculative.

6.116  Reduced enforcement costs. Third, copyright owners suggest that the 2010 Act will reduce their enforcement costs by providing a subsidised system for identifying and preventing wrongdoing.180 Their three suggested improvements are: first, a reduced need for litigation due to deterrent effects; second, where this prophylaxis proves insufficient, the availability of cheaper procedures to pursue primary wrongdoers; and third, a system for identifying the most egregious wrongdoers over time. The extent of any reduction in enforcement costs depends largely on the scheme’s effectiveness: if end-user litigation remains necessary, there is no reason to suppose that overall costs would be any lower; if anything, they may be higher due to the added delay and cost of sending notices.

6.117 Although ISPs face new obligations and potential liability under the 2010 Act, they benefit in three ways.

6.118  Standardisation of notices. First, the scheme offers standardised processes which reduce the overall costs of receiving and processing notifications. Because infringement reports are currently received in incompatible formats and media, ISPs must manually review thousands of daily notices; under the Code, reports will be submitted electronically in a standardised XML format.181 Ofcom will also supply a specification for matching IP addresses to subscribers and guidance on appropriate evidentiary standards. Following this guidance may reduce ISPs’ reputational and legal risk compared to self-regulation.

6.119  Lower risk of monetary liability. Second, the scheme reduces the pressure on ISPs and other internet services to supply monetary remedies for infringement. While participation does not confer immunity from authorisation liability—and therefore cannot be considered a safe harbour182—it is evidence of reasonable steps being taken to discourage infringement, which makes an inference of authorisation or joint wrongdoing less likely. Further, the scheme encourages cooperation between ISPs and claimants which may remove the impetus for litigation.183

6.120  Reduction in traffic costs. Finally, repeat infringers and infringing content disproportionately account for bandwidth utilisation.184 An effective graduated response system therefore promises, at least in theory, to reduce ISPs’ network traffic and operating costs, without necessarily impacting consumption patterns and revenues from non-infringing customers. More plausibly, the scheme may have the secondary effect of shifting users’ infringing network traffic to non-infringing sources (eg Netflix), resulting in the same or even higher bandwidth consumption. Some ISPs bundle internet access with lawful video and music services. Reducing infringement, it may be assumed, would benefit those services.

6.121  Fewer threats of proceedings. The 2010 Act offers six main benefits to internet users. First, because it encourages claimants to pursue targeted action rather than indiscriminate litigation against any suspected infringer, users are less likely to be threatened with legal action. Attention is focused on serial offenders, who have received prior warnings.185

6.122  Fair warning. Graduated response also avoids catching casual infringers unawares. Users accused of infringement once or twice might otherwise face a ‘mafia-like choice’ between costly settlement and costlier litigation.186

6.123  Education and supervision. Second, the scheme educates infringers—many of whom are likely to be minors—allowing households to put in place appropriate measures to self-regulate their children’s internet usage without immediate exposure to liability. Although often overlooked, this critically important function helps avoid the generation of ‘copyright criminals’ feared by Lessig.187

6.124  Indirect benefits. Third, reduced infringement levels may benefit non-infringers by increasing network performance, and the quality and variety of copyright materials.188

6.125  Reduced risk of unjustified threats. The Act and Code specify minimum standards for disclosure and basic processes for identification, which reduce the likelihood of users’ personal information being wrongfully disclosed to claimants. As a universal statutory scheme, rules are mandatory and identical as between ISPs, which promotes transparency and consistency.

6.126  Rights of appeal. Further, Ofcom and any appeal bodies are subject to judicial review and must act compatibly with human rights. These guarantees far exceed what a privately administered scheme would offer.

6.127 The 2010 Act also carries substantial economic, social, and technological disadvantages, which are analysed in the following paragraphs under the same headings.

6.128  Potentially limited effectiveness. The 2010 Act may not actually be effective in materially reducing infringement. Even if it is, claimants still bear the onus of detecting and notifying ISPs of alleged infringements. Currently, most repeat-players outsource these tasks to digital forensics firms, who employ automated monitoring tools to generate and send millions of notices each week. Because claimants must pay a per-notice fee, this will limit their ability to report infringements to ISPs. Even to maintain current levels of notifications would cost most claimants £5m annually,189 though this may serve to deter frivolous or unverified complaints.

6.129  Cost and delay. Second, the procedures for obtaining access to subscriber data are protracted. Unlike Norwich Pharmacal orders, which may (subject to proportionality considerations) be sought immediately from a single instance of prima facie infringement, the process for producing and disclosing infringer lists is lengthy and eventually requires claimants to go through the same process to obtain disclosure.

6.130  Reduced consumption. Finally, claimants may incur indirect costs from curtailing infringement. Strict enforcement may simply reduce the overall consumption of copyright works, thereby causing greater long-term harm to cultural industries.190

6.131  Higher operating costs. ISPs will incur various costs in establishing and operating the system: in particular, receiving reports; matching reports to subscribers; generating and sending notices; retaining records; compiling infringer lists; handling subscribers’ inquiries about notices; maintaining data security; and monitoring compliance.191

6.132 Technical measures would involve ‘far more burdensome technical obligations’192 and additional implementation and operational costs. All costs are highly sensitive to changes in the scheme’s efficacy:

Compliance cost figures are very sensitive to the underlying assumptions. If only 50% instead of 70% of infringers stopped, annual costs of compliance would increase from a range of £6–20 million to a range of £10–30 million. If instead of one letter [claimants] required two letters a year to be sent to serious infringers, the costs would double.193

6.133 To meet these variable costs, Ofcom has proposed a notification fee payable by each qualifying copyright owner equal to 75 per cent of the relevant costs.194 However, a fixed fee encourages incomplete investigation of complaints, and leaves ISPs with the variable risk of error. In the short term, making ISPs bear a percentage of costs may cause them to resist all efforts to make the scheme succeed or proceed to technical obligations, which is likely further to reduce its efficacy, potentially increasing overall costs.

6.134  Market distortion. ISPs have also criticised the quantitative threshold as distorting the market for internet access by encouraging infringers to shift to smaller or mobile ISPs who are exempt from Code obligations and able to offer lower-cost services. This issue was considered and dismissed by the Court of Appeal in BT, principally because there was no evidence this was likely to occur.195 Switching costs associated with early contractual termination and reconnection fees may inhibit the risk of make market distortions, especially given the price-sensitivity of infringers as a consumer group.

6.135  Higher internet access costs. The greatest costs are borne by internet users. First, users are likely to pay ISPs’ share of enforcement costs by way of higher access fees. This may widen the ‘digital divide and exclude marginal consumers from affordable broadband access’.196

6.136  False positives. Second, using IP address evidence will inevitably produce false-positive notifications which harm innocent subscribers’ interests. Flaws in IP address matching are widely acknowledged and have led to claims against inanimate printers,197 deceased grandmothers,198 children confined to hospital,199 and copyright owners themselves.200

6.137 There is an inherent tension between reducing enforcement costs for claimants and ensuring allegations of infringement are properly investigated, tested, and judicially reviewed.201 Glickman’s infamous comment summarises one extreme: ‘When you go trawling with a net, you catch a few dolphins.’202 Because reports under the 2010 Act are premised on mere assertions and honestly held beliefs, the Code does not come close to requiring ‘unequivocal and cogent evidence’ of the kind necessary to succeed in civil proceedings.203 Accordingly, it seems likely to catch some ‘dolphins’.

6.138 Although false-positive notifications do not necessarily cause any immediate harm to subscribers (since individually they create no liability), they do place them at a forensic disadvantage in subsequent infringement proceedings because they function as prima facie evidence of wrongdoing which may be difficult to rebut later. Nevertheless, repeated allegations against a subscriber are intrinsically different to repeated judicial determinations of infringement.

6.139 Possible solutions to the risk of false positives include increasing evidentiary standards; imposing penalties for reckless or dishonest allegations; requiring, by analogy with the rules in civil proceedings, claimants to bear all costs associated with an unsuccessful allegation;204 and accrediting ISPs’ and claimants’ detection processes, as occurs in France.205 However, because the ultimate question of infringement is not trivial to resolve,206 even these steps will not completely eliminate errors. Perplexingly, hard-core infringers are probably better off, while non-infringers are worse off if they are mistakenly targeted or forced to subsidise the costs of investigating infringers.

6.140  Reversed onus of proof. Third, the scheme reverses the onus of proof: alleged infringers are presumed tortfeasors, unlike in court proceedings.207 Subscribers must pay a £20 fee to appeal any notice, unless they succeed.208

6.141  Impact on ‘tolerated use’. Finally, the 2010 Act increases uncertainty about which informal dealings with copyright will continue to be tolerated. Many tortious but trivial infringements are not enforced by copyright owners, usually they cause minute damage, are prohibitively costly to prevent, or actually benefit the copyright owner.209 Katyal argues that there is a ‘pervasive divide’ between what copyright law requires and what the market tolerates.210 Graduated response does not distinguish between ‘technical’ infringements and more serious incursions; this threatens to prevent informal dealings with copyright which are tolerated and socially beneficial. The quantitative threshold is too low to accommodate these practices meaningfully.211

6.142 The concept of proportionality is discussed in detail in chapter 13.212 This section considers the proportionality of the scheme established under the 2010 Act, and graduated response schemes generally.

6.143 In BT, the Court concluded that the initial obligations were not disproportionate, a finding not disturbed on appeal.213 The Court was understandably slow to interfere in primary legislation, but Parker J’s approach to judicial review can only be described as ‘highly deferential’.214 In a scheme whose net effect hinges upon complex economic and behavioural evidence, this offers limited scope for challenging assumptions made by Parliament which, although wrong, are not irrational or wholly without basis.215

6.144 In seeking to reduce infringement and protect claimants’ property rights, the 2010 Act clearly pursues a legitimate aim. Those rights are recognised as fundamental rights.216 However, it is worth noting that this aim needs to be assessed in light of the wider objectives and policies of the copyright system: stronger exclusive rights encourage underutilisation, which carries social costs, as does spillover from excessive enforcement. Eventually, those costs will outweigh the benefits of stronger copyright protection.

6.145 Most objections to the notification scheme fall under one of four headings.

6.146  Effectiveness. First, the proportionality of the 2010 Act largely depends on how effective it is in achieving its objectives once implemented and, in particular, on its economic side effects.217 The French experience is revealing though ultimately inconclusive: of 736,000 notices sent, 62,000 subscribers have received a second notice and 165 have received a third.218 It is difficult to infer any deterrent effect from these data, since many notice recipients have reported obfuscating and continuing their infringing activity, though music sales and profits appear to have stabilised.219 Nevertheless, the ends likely to be achieved by the scheme are inconclusive, and will require close evaluation once implemented.

6.147  Innocent users. The second objection relates to subscribers who are themselves intermediaries—for example, where the primary infringements were committed by family members, guests, customers, or unauthorised users of a wireless network.220 It is correct that, absent something more, simply allowing someone to access a connection, or failing to secure it properly, is unlikely to be authorisation or joint tortfeasance. However, no direct liability attaches to being placed on an infringer list, and in a subsequent action for infringement the Court would apply ordinary liability rules (exonerating such a subscriber). Further, a statutory defence applies where a subscriber can show she was not the infringer and took ‘reasonable steps’ to secure her connection.221 This objection is therefore unconvincing, since the scheme does not expand tortious liability beyond existing limits.

6.148  ‘Chilling effects’. The third category of objections identifies possible ‘chilling effects’, such as subscribers refraining from non-tortious activities for fear of being placed on an infringer list,222 and access points—such as coffee shops, libraries, and hotels—ceasing to offer open access. In BT Parker J was not persuaded that these were credible prospects because claimants would still need to prove that the access provider had authorised infringement—‘a relatively high test’223—and so no disproportionate liability would result. Existing liability rules already apply to such intermediaries without difficulty, and the quantitative threshold in the Code is likely to exclude them in any case. Similarly, non-tortious dealings with copyright by subscribers seem unlikely to be deterred: exceptions and defences to infringement are recognised in subscribers’ grounds of appeal.224

6.149  Illegitimate subsidy. The fourth objection is to ISPs being required to subsidise the cost of enforcing private property rights at all. Under the scheme, ISPs must pay copyright owners despite making no recognised use of copyright material. This enlarges the statutory monopoly without corresponding social returns. Copyright owners respond that contributions are justified by ISPs’ facilitation of infringement and the need to encourage efficient compliance.225

6.150 Although a subsidy may be justified if ISPs benefited from wrongdoing, it is not obvious that this is so, since ISPs would often face comparable bandwidth costs if consumers sourced the same content from legitimate sources. The existence of enforcement duties does not necessarily mean ISPs should pay for those measures unless they are themselves secondary infringers. To require payment as a non-infringer is inconsistent with the approach taken to injunctions against other non-wrongdoers.

6.151  Notification duties. Despite these criticisms, it seems difficult to sustain an objection to notification being carried out by ISPs under the supervision of administrative and judicial bodies. Repeated allegations of infringement over a substantial period are unlikely all to be false-positives, and the effect of a notice is not even prima facie liability, but rather to educate and deter—even if only to a moderate degree.

6.152 Provided costs are appropriately borne by copyright owners rather than innocent consumers or service providers, soft measures of this kind are preferable to inflexible claims under primary and secondary liability rules. Although imperfect, the 2010 Act supplies adequate procedural guarantees to ensure that subscribers receive sufficient notice of allegations and time in which to read, act on, and, if necessary, contest them before incurring liability. The notification obligations contained in the scheme accordingly reflect a fair balance between claimants’, ISPs’, and consumers’ rights.

6.153  Technical measures. Technical measures would stand in a different category, since they are penalties that would disproportionately affect non-tortious uses of the internet. Although certain criminal penalties can prevent non-criminal activities,226 existing remedies for copyright infringement are mainly compensatory or restitutionary.227 Disconnection represents a different kind of remedy; it is not a transfer of value designed to compensate or disgorge the fruits of wrongdoing, but a punitive means of denying the defendant access to information—akin to digital incarceration.

6.154 In the terminology of layers, disconnection amounts to disabling the physical layer in order to regulate activity at the content layer. As Solum and Chung point out, this is inherently over-inclusive, since all communications—even those which do not fall within the regulatory purpose of disconnection—are prevented. Accepting such technical measures would necessarily be inconsistent with the end-to-end principle.228 Presumably for similar reasons, the High Court concluded in iiNet that disconnection would not be reasonable even if infringement had been proved, since it would disproportionately restrict lawful activities.229

6.155 Additionally, unlike notification, there is no necessary connection between the magnitude of infringement and the type of technical sanction applied. Minor home copying appears to count towards graduated response equally alongside commercial piracy and wilful self-enrichment. While any proportionate system of technical measures must align penalty with severity of wrongdoing, the 2010 Act contains no such guarantee.230

6.156 Indeed, the 2010 Act contains few details about when technical measures would be applied, for what duration, and by whom. Because their activation is a matter of largely ministerial discretion, ISPs could be held in a kind of regulatory purgatory under threat of technical measures. The Joint Committee on Human Rights criticised this ‘skeletal approach’ as inappropriate where powers could infringe users’ fundamental rights, since there is no prospect of parliamentary scrutiny of their exercise.231

6.157 Even assuming perfect detection and matching, there are doubts about how effectively technical measures would operate. Without coordination via some kind of centralised ‘blacklist’, subscribers could immediately reconnect with another ISP. There is nothing to prevent parallel infringement via smartphones, which are serviced by non-qualifying carriers. While a central subscriber registry would complete the analogy with driving demerit points, it would draw unwelcome comparisons with the identity registration requirements of autocratic states and only fuel concerns over ‘network surveillance’ and internet freedom.232 Ease of circumvention is not conclusive evidence of disproportionality, but it does reduce the ability of technical measures to achieve their stated aims.

6.158 In BT , Parker J did not assess the comparative effectiveness and intrusiveness of alternative solutions to online enforcement. However, there are several obvious candidates which may constitute less intrusive means of reducing infringement when compared to technical measures such as the disconnection of subscribers.

6.159  Removal of infringing material. First, notice-and-takedown regimes already operate effectively, though they suffer from cross-border limitations and need greater harmonisation.

6.160  Automated enforcement. Second, automated enforcement systems such as YouTube’s Content ID might provide greater accuracy and flexibility using fingerprinting technologies.

6.161  Injunctions. Third, more tailored injunctions—such as website blocking and de-indexing orders—233 could be sought to prevent access to the sources of infringing material. If effective, these orders would operate to prevent the substantial majority of subscribers to an ISP’s service from downloading the relevant material. This kind of prophylaxis may well be both more effective and less intrusive than technical measures directed at the consumer.

6.162  Other technical measures. Fourth, data caps, connection shaping (limiting a subscriber’s speed of access), or even sandboxing (restricting a subscriber to a whitelist of permitted news, entertainment, and government websites) could be just as effective as outright disconnection, with significantly fewer spillover effects.

6.163  Criminal sanctions. Fifth, technical measures could be reserved for those proved to be engaged in criminal infringements.234

6.164  Expedited claims. Sixth, graduated response could be abandoned in favour of an ‘expedited adjudication process’, as the OECD recommended,235 or fast track action against the primary wrongdoer.

6.165  Damages. Seventh, claimants could simply use existing remedies to claim damages from commercial infringers, distribution services, and bootleggers under existing procedures.

6.166  Positive incentives. Finally, as Patry observes, graduated response is ‘all stick and no carrot’.236 Regulators could reduce infringement by altering substantive copyright norms or encouraging claimants to make non-tortious modes of content acquisition more convenient, affordable, and accessible.

6.167 The availability of these alternative mechanisms suggest that, in combination with a notification and identity disclosure regime, other measures have the potential to be at least as effective as technical measures with fewer harms. It is regrettable that courts and policymakers have chosen not to pursue these alternatives.

Notes
1

Sam Ricketson, ‘Copyright’ in Blackshield, Coper, and Williams (eds), The Oxford Companion to the High Court of Australia (2001) 152, 154.

2

David Price, ‘Technical Report: An Estimate of Infringing Use of the Internet—Summary’ (Envisional Ltd, January 2011) <http://documents.envisional.com/docs/Envisional-Internet_Usage_Report-Summary.pdf>.

3

Viacom International Inc v YouTube Inc, Plaintiffs’ Memorandum, 8 (Case No 1:07-cv-02103, SDNY, 2012).

4

Robert Layton and Paul Watters, ‘Investigation into the Extent of Infringing Content on BitTorrent Networks’ (Internet Commerce Security Laboratory, April 2010) <http://afact.org.au/research/bt_report_final.pdf>.

5

Sandvine Inc, Global Internet Phenomena Report (2014) 6 (file-sharing traffic declined from 31 per cent to 8.3 per cent of traffic between 2008 and the first half of 2014).

6

Stevens v Kabushiki Kaisha Sony Computer Entertainment (2005) 224 CLR 193, 207–8 (Gleeson CJ, Gummow, Hayne, and Heydon JJ).

7

For detailed coverage of the scope of primary liability, readers are referred to a specialist work: see, eg, Mary Vitoria et al, Laddie, Prescott and Vitoria: The Modern Law of Copyright and Designs (4th ed, 2011) chs 14–19.

8

Donaldson v Becket (1776) 1 ER 835, 839. See now Copyright, Designs and Patents Act 1988 s 171(2).

9

(‘1988 Act’) (emphasis added).

10

1988 Act ss 96(1), 101(2).

11

1988 Act ss 16(1)(a), 17(2).

12

1988 Act ss 16(1)(d), 20(2).

13

For discussion of the right of communication to the public in the context of internet intermediaries, see Twentieth Century Fox Film Corporation v Sky UK Ltd [2015] EWHC 1082 (Ch), [32]–[42] (Birss J) (website operator); Svensson v Retriever Sverige AB, C-466/12, EU:C:2014:76 [2014] Bus LR 259, [15]–[23] (gateway); Newspaper Licensing Agency Ltd v Meltwater Holding BV, C-360/13, EU:C:2014:1195; [2014] AC 1438, [57] (website operator); and ITV Broadcasting Ltd v TVCatchup Ltd [No 2], C-607/11, EU:C:2013:147 [2013] Bus LR 1020, [21], [31] (platform).

14

See, eg, Tilman Lüder, ‘The Next Ten Years in EU Copyright: Making Markets Work’ (2007) 18 Fordham Intellectual Property, Media & Entertainment Law Journal 1, 21–8; Bernt Hugenholtz et al, The Recasting of Copyright & Related Rights for the Knowledge Economy (2006).

15

Eg, if a website operator selects and uploads images or videos which involve unauthorised copying, the acts giving rise to liability are committed by that operator alone.

16

See, eg, 1988 Act ss 23–26, 184(1)(b), 296ZG(2).

17

The phrase was first used in the Australian Copyright Act 1905 (Cth) s 13(1) (which described the right to ‘authorize another person to do’ the acts comprising copyright). See also ss 14(1) (dramatic and musical works), 34 (artistic works).

18

Copyright Act 1911 (Imp) s 1(2) (‘1911 Act’).

19

Jane Ginsburg and Sam Ricketson, ‘Inducers and Authorisers: A Comparison of the US Supreme Court’s Grokster Decision and the Australian Federal Court’s KaZaA Ruling’ (2006) Columbia Public Law & Legal Theory Working Papers, Paper 0698, 10.

20

Robert Burrell, ‘Copyright Reform in the Early Twentieth Century: The View from Australia’ (2006) 27 Journal of Legal History 239, 239.

21

WEA International Inc v Hanimex Corporation Ltd (1987) 17 FCR 274, 285 (Gummow J) (‘WEA’).

22

See, eg, Dramatic Literary Property Act 1833 (3 & 4 Will IV, c 15) s 2 (infringement by one who shall ‘represent or cause to be represented’ a work without consent); Copyright Act 1842 (5 & 6 Vict, c 45) s 21.

23

[1849] 8 CBR 836; 137 ER 737, 742.

24

Karno v Pathé Frères (1909) 100 LT 260, 262 (Vaughan Williams J) (‘Karno’).

25

Falcon, 491 (Bankes LJ), 496 (Scrutton LJ).

26

Walter Copinger, The Law of Copyright (5th ed, 1915) 136.

27

See, eg, E J Macgillivray, The Copyright Act, 1911: Annotated (1912) 22 (arguing that the references to authorisation ‘appear to be superfluous’ and ‘unnecessary’).

28

[1924] 1 KB 1, 10–11 (Bankes LJ), 12 (Scrutton LJ), 15 (Atkin LJ) (‘Ciryl’).

29

See Performing Right Society Ltd v Mitchell and Booker (Palais de Danse) Ltd [1924] 1 KB 762, 773 (McCardie J) (‘Mitchell’).

30

Falcon v The Famous Players Film Co Ltd [1926] 2 KB 474 (‘Falcon’).

31

Falcon, 491 (Bankes LJ), 496–7 (Scrutton LJ) (disposing of the appeal on the basis of primary infringement).

32

Falcon, 491 (Bankes LJ).

33

Falcon, 499 (Atkin LJ).

34

CBS Inc v Ames Records & Tapes Ltd [1982] Ch 91, 110 (Whitford J).

35

Mitchell, 765 (McCardie J).

36

See Hugh Laddie, Peter Prescott, and Mary Vitoria, The Modern Law of Copyright and Designs (2nd ed, 1995) 911 (the formula ‘countenance, sanction, approve’ replaces ‘one vague expression by another equally vague one’).

37

See, eg, Evans v Hulton & Co (1924) 131 LT 534, 535 (Tomlin J); Monckton v Pathé Frères Pathephone Ltd [1914] 1 KB 395, 403 (Buckley LJ); A & M Records, 9–10 (Foster J) (countenance, sanction, and approve). cf Ames, 106 (Whitford J) (grant of a right).

38

Ash v Hutchinson and Co (Publishers) Ltd [1936] 1 Ch 489, 497 (Slesser LJ), 502 (Romer LJ), 506–7 (Greene LJ) (‘Ash’).

39

Mitchell, 773 (McCardie J).

40

A & M Records Inc v Audio Magnetics Incorporated (UK) Ltd [1979] FSR 1 (‘A&M Records’).

41

A&M Records, 9–10 (Foster J).

42

CBS Inc v Ames Records & Tapes Ltd [1982] Ch 91, 99 (Whitford J) (‘Ames’).

43

Ames, 118. The claimants relied on evidence from private investigators who conducted ‘trap’ purchases: at 104. Cf RCA Records v All-Fast Systems Inc, 594 F Supp 335 (1984) (where the defendant also supplied the retaping machine on its premises).

44

Cf The University of New South Wales v Moorhouse (1975) 133 CLR 1, 13 (Gibbs J), 23 (Jacobs J).

45

CBS Songs Ltd v Amstrad Consumer Electronics plc [1988] 1 AC 1013 (‘CBS’).

46

As to the former, see chapter 5, section 2. See also Fish & Fish Ltd v Sea Shepherd UK [2015] AC 1229, [41] (Lord Sumption JSC).

47

CBS, 1050 (Lord Templeman).

48

Amstrad Consumer Electronics plc v The British Phonographic Industry Ltd [1986] FSR 159, 171–2, 186–7 (Whitford J) (‘Amstrad v BPI’).

49

CBS, 189 (Whitford J).

50

CBS, 207 (Lawton LJ), 211 (Slade LJ), 217–18 (Glidewell LJ); CBS, 1054–5 (Lord Templeman) (Lord Keith, Lord Griffiths, Lord Oliver, and Lord Jauncey agreeing).

51

Amstrad v BPI, 213–14 (Slade LJ), 219 (Glidewell LJ). The reasoning in Ames appears to have assumed that no action lay for negligently facilitating a third party’s infringement of copyright: see Paterson Zochonis Ltd v Marfarken Packaging Ltd [1983] FSR 273, 296 (Robert Goff LJ) (‘Paterson Zochonis’).

52

See Dorset, 1030 (Lord Reid), 1060 (Lord Diplock).

53

Paterson Zochonis, 299 (Robert Goff LJ).

54

Paterson Zochonis, 285 (Oliver LJ), 288 (Fox LJ).

55

Paterson Zochonis, 290 (Fox LJ), 301 (Robert Goff LJ).

56

Cf Caparo Industries plc v Dickman [1990] 2 AC 605, 634, 642–3 (Lord Oliver) (Lord Ackner agreeing), 655 (Lord Jauncey); Muirhead v Industrial Tank Specialities Ltd [1986] QB 507, 528–9 (Robert Goff LJ), 534 (Nourse LJ).

57

Amstrad v BPI, 214 (Slade LJ).

58

1988 Act ss 16(1) (copyright the right to do specified acts ‘in the United Kingdom’).

59

[1995] RPC 657 (‘ABKCO’).

60

See Clark (Inspector of Taxes) v Oceanic Contractors Inc [1983] 2 AC 130, 144–5 (Lord Scarman).

61

ABKCO, 660 (Hoffmann LJ).

62

[2010] EWHC 2911 (Ch), [28]–[30] (Floyd J); rev’d [2011] 1 WLR 3044 on other grounds.

63

WEA, 283 (Gummow J).

64

CBS, 1058E (Lord Templeman); MCA Records, 424 (Chadwick LJ). See chapter 5, section 2.

65

Unilever, 608 (Mustill LJ); MCA Records, 418 (Chadwick LJ); Belegging, 66 (Buckley LJ); Lumley, 213 (Erle J).

66

See Roadshow Films Pty Ltd v iiNet Ltd (2012) 248 CLR 42, [100] (Gummow and Hayne JJ).

67

See Coin Controls Ltd v Suzo International (UK) Ltd [1999] Ch 33, 39–40 (Laddie J) (‘Coin Controls’).

68

Coin Controls, 40–1 (Laddie J).

69

Newzbin, [103] (Kitchin J). See also Ash, 508 (Greene LJ); Unilever, 603 (Mustill LJ) (Ralph Gibson and Slade LJJ agreeing). See chapter 5, section 2.1.

70

CBS, 1058 (Lord Templeman).

71

See, eg, Metro-Goldwyn-Mayer Studios Inc v Grokster Ltd, 545 US 913, 935–6 (2005).

72

‘Cause’ should be understood in the sense of belonging to a set of conduct sufficient for the primary infringement to occur: see, by analogy, Richard Wright, ‘The NESS Account of Natural Causation: A Response to Criticisms’ in R Goldberg (ed), Perspectives on Causation (2011) 285, 291. Cf Herbert Hart and Tony Honoré, Causation in the Law (2nd ed, 1985) 112; Tony Honoré, ‘Necessary and Sufficient Conditions in Tort Law’ in D G Owen (ed), Philosophical Foundations of Tort Law (1990) 363.

73

A similar approach prevails in the United States: see Viacom v YouTube, 15; Perfect 10 Inc v CCBill LLC, 488 F 3d 1102 (9th Cir, 2007); Corbis Corp v Amazon.com Inc, 351 F Supp 2d 1090, 1108 (WD Wash, 2004).

74

Amstrad v BPI, 212–13 (Slade LJ), 219 (Glidewell LJ). See also Dunlop Pneumatic Tyre Co Ltd v David Moseley & Sons Ltd [1904] 1 Ch 612, 620 (Stirling LJ), 621 (Cozens-Hardy LJ).

75

Amstrad v BPI, 206 (Lawton LJ).

76

Unilever, 608 (Mustill LJ).

77

(1863) 3 B & S 556, 563; 122 ER 209, 213 (Cockburn CJ) (‘Lyon’); aff’d (1864) 5 B & S 751; 122 ER 1010.

78

Macaulay v Screenkarn Ltd [1987] FSR 257, 259 (Falconer J).

79

Grower v British Broadcasting Corporation [1990] FSR 595, 611 (Knox J). Logically, if a contractual provision prohibits the infringing acts, then—assuming it is a genuine term and not mere window dressing—this should be one factor militating against the existence of a common design.

80

CBS, 1055, 1057 (Lord Templeman) (Lord Keith, Lord Griffiths, Lord Oliver, and Lord Jauncey agreeing).

81

Magical Marketing Ltd v Holly [2009] ECC 10, [60] (Norris J).

82

See Douglas v Hello! Ltd [No 2] [2003] EMLR 28, 598 (Rix LJ) (‘Douglas [No 2]’).

83

Douglas [No 2], 598 (Rix LJ).

84

CBS Songs Ltd v Amstrad Consumer Electronics plc [1988] Ch 61, 78 (Nicholls LJ) (Fox LJ agreeing).

85

iiNet, [55] (French CJ, Crennan and Kiefel JJ).

86

Randal Picker, ‘Copyright as Entry Policy: The Case of Digital’ (2002) 47 Antitrust Bulletin 423, 442; cited in In re Aimster Copyright Litigation, 334 F 3d 643, 645 (7th Cir, 2003) (Posner J).

87

Twentieth Century Fox Film Corp v Newzbin Ltd [2010] EWHC 608 (Ch); [2010] FSR 21 (‘Newzbin’).

88

Newzbin, 522 (Kitchin J).

89

Newzbin, 526, 537 (Kitchin J).

90

Newzbin, 540 (Kitchin J).

91

Newzbin, 543 (Kitchin J) (emphasis added).

92

Newzbin, 545–6 (Kitchin J).

93

See Simon Baggs and Rachel Barber, ‘A Changing Tide in the Fight against Online Piracy: How Significant is the Newzbin Judgment?’ [2010] Entertainment Law Review 234, 237.

94

Newzbin, 545–6 (Kitchin J).

95

Dramatico Entertainment Ltd v British Sky Broadcasting Ltd [2012] EWHC 268 (Ch), [81] (Arnold J) (‘Dramatico’).

96

Dramatico, [75]–[77] (Arnold J).

97

Dramatico, [81] (Arnold J).

98

Dramatico, [83] (Arnold J).

99

Universal Music Australia Pty Ltd v Cooper (2005) 150 FCR 1 (‘Cooper’).

100

Cooper v Universal Music Australia Pty Ltd (2007) 156 FCR 380, 389 (Branson J).

101

Cooper, 30–1 (Tamberlin J); aff’d (2007) 156 FCR 380, 393 (Branson J) (French J agreeing).

102

Lacoste v Multimania Production SA [2001] ECC 22, 204 (a French case preceding the Directive). See further chapter 12, section 5.

103

Cooper, 20–1 (Tamberlin J).

104

Cf Roadshow Films Pty Ltd v iiNet Ltd (2010) 263 ALR 21, [398] (Cowdroy J); rev’d (2011) 194 FCR 285.

105

Cooper, 29 (Tamberlin J).

106

(2010) 263 ALR 21 (‘iiNet (Trial)’).

107

Roadshow Films Pty Ltd v iiNet Ltd (2012) 248 CLR 42, [59] (French CJ, Crennan and Kiefel JJ) (‘iiNet’).

108

iiNet, [115] (Gummow and Hayne JJ).

109

iiNet (Trial), [400]–[402] (Cowdroy J).

110

iiNet, [38].

111

iiNet, [18], [38], [64] (French CJ, Crennan and Kiefel JJ). See also iiNet (Trial), [411] (Cowdroy J).

112

iiNet, [146] (Gummow and Hayne JJ).

113

iiNet, [19]–[20], [65] (French CJ, Crennan and Kiefel JJ).

114

iiNet, [115]–[116] (Gummow and Hayne JJ).

115

iiNet, [146] (Gummow and Hayne JJ).

116

iiNet, [75]–[78] (French CJ, Crennan and Kiefel JJ), [146] (Gummow and Hayne JJ).

117

iiNet, [127] (Gummow and Hayne JJ).

118

iiNet, [125] (Gummow and Hayne JJ). Moreover, the statutory factors specified in s 101(1A) of the Copyright Act 1968 (Cth) suggested a narrower inquiry.

119

iiNet, [67] (French CJ, Crennan and Kiefel JJ).

120

iiNet, [69] (French CJ, Crennan and Kiefel JJ); citing WEA, 286–8.

121

Newzbin, [95] (Kitchin J).

122

Case C-484/14, McFadden [2015] OJ C46/18. See chapter 12, section 3.

123

David Lindsay, ‘Liability of ISPs for End-User Copyright Infringement’ (2010) 29 Communications Law Review 1, 12.

124

See Sony Music Entertainment (UK) Ltd v Easyinternetcafe Ltd [2003] IP&T 1059, 1066 (Peter Smith J).

125

Ciryl, 9 (Bankes LJ).

126

CBS, 1054.

127

[2012] EWHC 1185 (Ch) (‘Sportradar’).

128

Sportradar, [84], [81] (Floyd J).

129

Perfect 10 Inc v Visa International Service Association, 494 F 3d 788, 796 (9th Cir, 2007) (‘Perfect 10’).

130

Perfect 10, 797–8.

131

See generally Peter Yu, ‘Digital Copyright and Confuzzling Rhetoric’ (2011) 13 Vanderbilt Journal of Entertainment and Technology Law 881.

132

Yana Breindl and François Briatte, ‘Digital Network Repertoires and the Contentious Politics of Digital Copyright in France and the European Union’ (Paper presented at the Oxford Internet Institute conference, Oxford, 16 September 2010) 6.

133

Nicolas Suzor and Brian Fitzgerald, ‘The Legitimacy of Graduated Response Schemes in Copyright Law’ (2011) 34 University of New South Wales Law Journal 1, 1.

134

Lawrence Freedman, Kennedy’s Wars: Berlin, Cuba, Laos, and Vietnam (2002) 93–4.

135

Peter Yu, ‘The Graduated Response’ (2010) 62 Florida Law Review 1373, 1400.

136

See, eg, Gowers Review, [5.92]–[5.100]. cf Creative Britain, 50. See also BERR, ‘Consultation Document on Legislative Options to Address Illicit Peer-to-Peer (P2P) File-Sharing’ (July 2008) 47.

137

Ofcom, Online Copyright Infringement Initial Obligations Code (2010) (‘Code’).

138

Communications Act 2003 (UK) s 124D(1) (‘Communications Act’).

139

Online Infringement of Copyright (Initial Obligations) Cost Sharing Order 2010 (UK).

140

Communications Act s 124C(5); Code § 2.4.2; 2010 Act s 124N.

141

The main reasons are: (1) the higher cost of bandwidth to consumers; (2) ad valorem usage pricing; (3) latency and speed limitations; (4) network traffic management.

142

Most mobile networks assign addresses which may be shared between multiple subscribers, making identification much more difficult and unreliable.

143

Ofcom, Online Infringement of Copyright and the Digital Economy Act 2010: Notice of Ofcom’s Proposal to Make by Order a Code for Regulating the Initial Obligations (26 June 2012) 33 (‘Notice’).