Landmark ISP liability case decided in Australia

What did you say Skippy? ISPs are not liable for the infringement committed by their customers?

As was mentioned last week, we have been expecting an important ruling with regards to internet service provider (ISP) liability from Australia.  Behold Roadshow Films Pty Ltd v iiNet Limited [2010] FCA 24.  This is a case of tremendous importance because it is one of the first skirmishes in the brewing struggle between content owners and intermediaries (I use the term intermediaries on purpose, as the wider dispute encompasses companies such as Google).

iiNet is an Australian internet provider, which was sued by Australian film producer Roadshow Films, part of the Village Roadshow conglomerate.  However, the real power behind the suit was the weight provided by several Hollywood studios, including Warner, Columbia, 20th Century, and Sony Pictures.  The question at the heart of the proceedings was whether an ISP is to be held liable for the copyright infringement committed by its customers.  Call me cynical, but it seems like the film studios chose Australia as the first place to test this theory because Australian copyright law holds secondary infringement as direct infringement.  In other words, anyone making available copyright works will be held similarly liable, whether they do so directly or indirectly.

This is a lengthy and complex ruling, but it is remarkable that it has fallen upon a judge that seems to get the importance of the ruling in the wider context, and also who got the technical complexities involved.  Cowdroy J has managed to wade through the technical issues with exceptional clarity, and has produced a ruling that should become an instant classic.  The judge accurately identifies that the case hinges on two simple questions.  Have the iiNet customers infringed copyright directly?  The answer is yes.  Has iiNet authorised the copyright infringement of its users by failing to take steps to stop it from happening? Here the answer is no. The reasoning behind the negative answer is set out clearly:

“12. Firstly, in the law of authorisation, there is a distinction to be drawn between the provision of the ‘means’ of infringement compared to the provision of a precondition to infringement occurring. The decisions in Moorhouse, Jain, Metro, Cooper and Kazaa are each examples of cases in which the authorisers provided the ‘means’ of infringement. But, unlike those decisions, I find that the mere provision of access to the internet is not the ‘means’ of infringement. There does not appear to be any way to infringe the applicants’ copyright from the mere use of the internet. Rather, the ‘means’ by which the applicants’ copyright is infringed is an iiNet user’s use of the constituent parts of the BitTorrent system. iiNet has no control over the BitTorrent system and is not responsible for the operation of the BitTorrent system.
13. Secondly, I find that a scheme for notification, suspension and termination of customer accounts is not, in this instance, a relevant power to prevent copyright infringement pursuant to s 101(1A) […]
14. Thirdly, I find that iiNet simply cannot be seen as sanctioning, approving or countenancing copyright infringement. The requisite element of favouring infringement on the evidence simply does not exist. The evidence establishes that iiNet has done no more than to provide an internet service to its users. This can be clearly contrasted with the respondents in the Cooper and Kazaa proceedings, in which the respondents intended copyright infringements to occur, and in circumstances where the website and software respectively were deliberately structured to achieve this result.
15. Consequently, I find that the applicants’ Amended Application before me must fail.”

This conclusion is both astute and well informed, as it takes into account the very clear distinction between an internet service provider, and the producer of software designed specifically to authorise copyright infringement, such as Kazaa.  Cowdroy J goes in detail through some of the most relevant aspects of the technology, namely the intricacies of Bittorrent transactions, and also whether the identification of customers is reliable.  To me one of the most important parts of the ruling is that we get a very clear legal explanation of the Bittorrent protocol:

“70. To use the rather colourful imagery that internet piracy conjures up in a highly imperfect analogy, the file being shared in the swarm is the treasure, the BitTorrent client is the ship, the .torrent file is the treasure map, The Pirate Bay provides treasure maps free of charge and the tracker is the wise old man that needs to be consulted to understand the treasure map.
71. Whilst such an analogy grossly oversimplifies the situation it will suffice for present purposes. It demonstrates that all of the constituent parts of the BitTorrent protocol must work together before a person can access the file sought. In this judgment the Court will refer to all the constituent parts together as the ‘BitTorrent system’.
72. Such analogy also demonstrates that a number of deliberate steps are required to be taken by a person to bring about the means to infringe the applicants’ copyright. The person must download a BitTorrent client like Vuze, seek out .torrent files related to copyright material from websites, and download those .torrent files and open them in their BitTorrent client. Thereafter, the person must maintain connection to the internet for as long as is necessary to download all the pieces. The length of this downloading process will depend on the size of the file, the number of peers in the swarm and the speed of those peers’ internet connections.”

The other remarkable part of the ruling is that we get a very detailed insight into the technology involved in the detection of repeat infringers.  The applicants use a Bittorrent client called DtecNet, which singles out infringing torrents identified by the content owners, and starts downloading from the participant sharers in the swarm.  The agent downloads a full copy of the file to establish that this is indeed an infringing copy, and then identifies IP addresses in the iiNet network.  The agent would then download one piece of the full file from a specific IP address, and would continue to download a piece every 24 hours from the same address.  This technical operation was able to establish that there were indeed iiNet customers sharing infringing files repeatedly, but here is where one of the most interesting findings takes place.  Cowdroy J accurately points out that the infringement is much less widespread than previously beieved, and particularly, seemed to express doubt as to the efficiency of matching IP addresses with specific customers.  Another interesting technical issue that came to bear on the liability issue was the fact that the applicants were trying to paint a picture that a large portion of an infringer’s broadband use would be dedicated to illegal filesharing, while the evidence was to the contrary.

Turning to the legal aspects, one of the most important questions in the ruling is whether each time a user connects to the internet it should be considered as a fresh count of making the work available to the public, or whether the first action, namely downloading the .torrent file and seeding it, would be one single count.  Cowdroy J opines:

“310. [..] The Court finds that it is the wrong approach to focus on each individual piece of the file transmitted within the swarm as an individual example of an ‘electronic transmission’. The BitTorrent system does not exist outside of the aggregate effect of those transmissions, since a person seeks the whole of the file, not a piece of it. In short, BitTorrent is not the individual transmissions, it is the swarm. It is absurd to suggest that since the applicants’ evidence only demonstrates that one piece of a file has been downloaded by the DtecNet Agent from each iiNet user (in some cases more than one, but not many more), the applicants cannot prove that there have been ‘electronic transmissions’ by iiNet users of the applicants’ films. But it is equally absurd to suggest that each and every piece taken by the DtecNet Agent from an iiNet user constitutes an individual ‘electronic transmission’ infringement.
311. The correct approach is to view the swarm as an entity in itself. The ‘electronic transmission’ act occurs between the iiNet user/peer and the swarm, not between each individual peer. One-on-one communications between peers is the technical process by which the data is transferred, but that does not mean that such level of detail is necessarily what the communication right in s 86(c) focuses upon. While the DtecNet evidence cannot prove directly that an iiNet user has ‘electronically transmitted’ a film to the swarm (it can only show that the data has been ‘electronically transmitted’ to the DtecNet Agent acting as a peer in the swarm) the evidence is sufficient to draw an inference that in most cases iiNet users have done so. “

There are many other legal aspect treated, most of them are specific to Australian copyright law.  The most relevant from an international perspective, and one that has been highlighted elsewhere, is that the ruling seems to seriously attack the viability of three-strikes approaches to copyright infringement.  Firstly, there is a lengthy discussion about the legal reliability of evidence obtained by agents acting on behalf of the copyright owners, when by definition those agents are licensed by the owner to undertake actions that would otherwise be infringing, namely, collect copies from the internet.  Secondly, there is the troublesome issue of how such evidence should be suspect in the first place.   Cowdroy J says:

“631. […] the Court finds that it would not be appropriate to construe the safe harbour provisions such that there is an expectation on the [ISP] to terminate its subscribers at the request of a person who does not swear to the truth of his statement, and is an employee of an organisation whose precise legal status vis-à-vis the relevant copyright owners and exclusive licensees is not at all clear. Allegations of copyright infringement are serious charges which are potentially defamatory. Further, AFACT (the Australian copyright enforcement agency) enjoys no status as an authority invested with power to issue legally enforceable directions.”

This paragraph should be printed out and handed out to the Lords currently discussing the Digital Economy Bill (wink, wink, ORG).  We cannot possibly build a system of punishment that relies entirely on suspect accusations without any legal recourse.  The very principles of fairness and due process of law are at stake here.

What a refreshing ruling from Down Under.  Unfortunately, another copyright case from the land where women glow and men plunder has overshadow this more important development.

I just smiled and gave me a vegamite sandwich.