As I anticipated on Friday, and noted earlier today: the Full Federal Court judgment in the Cooper case has been handed down. This is a significant, appeal judgment on the scope of secondary liability in copyright law: that is, when can one person be responsible for the infringing activities of another?

In this case, the Full Federal Court had to consider whether a website, structured and designed both to provide links to infringing MP3 files, and to provide facilities for the easy, automatic upload of such links, could lead to liability for authorising infringement of copyright for the website designer (Cooper), and the hosting ISP (E-Talk) (the website itself, now down, can still be seen (though not used) via the Internet Archive Way Back Machine (the URL was To be completely clear, Cooper was not hosting infringing mp3s. But his website did provide easy access and a central point for placing links.

At first instance, Tamberlin J held both liable – leading to much commentary, particularly on the breadth of the Australian concept of authorisation. If you’re interested in where Australian law on secondary liability for copyright infringement lay prior to this judgment, and how it compares to US concepts, have a read of this article by copyright leading lights Sam Ricketson (australia) and Jane Ginsburg (US).

Summary: in this judgment, the Full Federal Court has pretty much affirmed the reasoning of the Trial Judge. Overall, I find the reasoning pretty troubling in this case: particularly the reasoning of Branson J, which seems to me to endorse a broader view of appropriate liability than the other judgment of Kenny J. I think the judgment shows three things:

  1. That Australian law is out on its own in terms of potential liability for authorisation of copyright infringement. The law is certainly broader – that is, the scope of activities that will potentially lead to liability is wider – than equivalent concepts in the UK (as illustrated in cases like the Amstrad case) or Canada (as illustrated by cases like CCH). And, as Ricketson and Ginsburg point out, it is broader, even, than US law post-Grokster. Anything that would be caught by post-Grokster inducement liability would also be caught by Australian authorisation liability – and then some.
  2. More broadly, the case illustrates that Australian copyright law is increasingly becoming a strange, independent beast. The judgments in this case are truly remarkable for their lack of reference to, or engagement with, recent overseas authority or legal developments. Just remarkable.
  3. Most remarkably to anyone outside that arcane society of the High Priests and Initiates of Copyright: linking to another website that carries an infringing file does carry some legal risk under the reasoning in this case.

Over the fold, I have some more detail. But let me give you a flavour of what I think. As I read the judgment, quite honestly, I was amazed by the absence of any explicit conscious engagement with the real world. To read this judgment – as I said, particularly the broader one of Branson J – you could be forgive for wondering how much the honourable members of the court actually use the Internet. It’s not way the law is described, it is the fact that the judgments can make comments about the potential liability of a website operator for copyright infringement for the most mundane acts – like, oh – linking to another website – without any acknowledgment that that might be an issue, or a problem, or an even slightly undesirable development in the law.

Here’s one favourite quote that captures some of the flavour of the Branson judgment:

‘Mr Cooper placed considerable weight on a suggested analogy between his website and Google. Two things may be said in this regard. First, Mr Cooper’s assumption that Google’s activities in Australia do not result in infringements of the Act is untested.’

This quote is striking. Not because it is inaccurate. On its terms, it is obviously perfectly true. No one has sued Google here yet. What is striking is that a statement as potentially momentous as this: that the activity of running a search engine – one of the fundamental activities that makes the ‘Internet work’ these days could well be infringement, we don’t know – can be said without the merest bat of a judicial eyelash.

Have we, or have we not just had a very extended debate about copyright law in Australia? Was not one of the memes in that debate the idea that copyright ought to ‘work’ in a digital environment? Are not search engines, and links, fundamental to the way the Internet and digital environment work? Did all this debate completely pass the members of the court by?

Now, you might say – well, naturally, the job of the court is to interpret the law, and all that ‘debate’ and even how the Internet generally operates is not relevant : the court just has to say the law ‘as it is’, bound, as it is, by precedent including High Court precedent. To some extent, certainly. But there really is more to it than that. The law of authorisation is essentially a common law doctrine, developed through the courts. The meaning of ‘to authorise’ is a subject of considerable debate. There are real differences between the judgments of the High Court in cases like Moorhouse. There is capacity, in this caselaw, and even in the relevant statutory provisions, to take into account the circumstances of the industry, industry practice, etc. It is not ok, in my view, to ignore all that. It means that the judgment takes on an air of unreality, and we get quotes, in the course of the reasoning, that I can just see being repeated in nasty ‘cease and desist’ letters all over the land.

I should be clear. It is not necessarily my view that Cooper himself should have got off scot-free. He did more than just set up a few links to a few infringing items. It is however my view that the idea of making someone liable for authorising infringement by linking to other websites is something we should take considerable care with.

My issue is that the judgment of Branson J in particular is stated in broad, unqualified terms that scarcely acknowledge the existence of another view or the need to exercise care in imposing liabilith. The judgment of Justice Kenny is less unqualified, and for that reason, less troubling. But even here there are some issues.

OK, so now I’ve got those immediate thoughts of my chest, and given a broad summary, let’s dive into a little of the detail of this judgment – and a little preliminary comparison of the two main judgments – of Kenny J and Branson J (French J apparently agrees with both, which doesn’t help us work out which set of reasoning is more likely to get built on in later cases).

Now, in broad terms, the law is:

  1. a person can be liable both for infringing copyright, and for authorising infringement (ss36, 101)
  2. ‘authorise’ has a meaning developed in the caselaw – basically, ‘sanction, coutenance or approve’. The classic old case on this was Moorhouse, in which a university was held liable because it provided coin-operated photocopiers right beside the books with no control and no steps to limit infringement;
  3. according to legislation introduced in 2000 and designed, allegedly, to codify common law and provide ‘certainty’, in assessing authorisation, courts must take into account three things:
  • the extent (if any) of the person’s power to prevent the doing of the act concerned
  • the nature of any relationship existing between the person and the infringer;
  • whether the person took any other reasonable steps to prevent or avoid the infringement.

Now, there’s been all kinds of debate over this kind of liability. How much ‘control’ is necessary over the infringers’ actions? How much knowledge? How close a relationship? Broadly, what the courts have done (in cases like the Kazaa case) is say it’s a case-by-case analysis. We have, in the existing caselaw, little by way of real guidelines which can help a person avoid infringement. I’ve complained about this before (start here and scroll up – this is my commentary back when Kazaa was handed down).Here’s what the judges in Cooper had to say:

Power to prevent infringement

Cooper pointed out, in the case, that he didn’t have power to prevent particular infringements occurring (in an argument somewhat reminiscent that accepted by the US 9th Circuit Court of Appeals in the Grokster case). That is because he set up a website with links. He didn’t have control over what people did with it – what links were submitted, what links were clicked on, what transmissions occurred as a result of such clicking on links.

Now, in a sense, this was never all that appealing on the face of it – mainly because Cooper did set up a website to take links, automatically, and never exercised any control over what links were up there. It was designed to have links to MP3s.

What is striking, however, is the breadth of Branson J’s comments on this. As she puts it:

‘I conclude that, within the meaning of the paragraph, a person’s power to prevent the doing of an act comprised in a copyright includes the person’s power not to facilitate the doing of that act by, for example, making available to the public a technical capacity calculated to lead to the doing of that act. The evidence leads to the inexorable inference that it was the deliberate choice of Mr Cooper to establish and maintain his website in a form which did not give him the power immediately to prevent, or immediately to restrict, internet users from using links on his website to access remote websites for the purpose of copying sound recordings in which copyright subsisted.’

In other words, it would seem – Cooper had the ‘power to prevent’ infringements because he had the power to take the website down, or to design it a different way.

Now, how far are we meant to take this? What if you design some software? Or a piece of hardware, like an MP3 player? You want to sell the player, and do. it turns out it is used for infringements in the marketplace, and that’s what people are buying it for. Presumably, your ‘power to prevent’ includes here a power ‘not to facilitate by making available to the public a technical capacity (the player, the software) calculated to lead to the doing of the infringing act.

Those of you who are familiar with the caselaw in this area may remember that one of the issues, in Grokster in the US, that led to the 9th Circuit holding Grokster not liable, was the fact that they did not control, or have the power to prevent, individual infringements. One thing that the US Supreme Court did was ‘invent’ a form of liability – inducement liability – that could lead to liability where a person was encouraged, induced etc to infringe – even where the individual acts were not controlled.

Look again at that Branson J reasoning. She finds sufficient ‘power to prevent’ in ‘making available to the public a technical capacity calculated to lead to the doing of that act’. I find that troubling.

Kenny J’s judgment is slightly less broad, I think. She does two things differently. First, she emphasises, far more than Branson J does, the point that it is the combination of factors – a combination, for example, of inactivity, knowledge, and control, that can lead to authorisation liability, depending on all the circumstance of the case. The critical role of the combination – that the whole can be greater than the sum of its parts – is lost in Branson J’s less sophisticated discussion. Second, Kenny J places more emphasis on the fact that Cooper could be seen to be encouraging infringements:

‘Every time an internet user activated a link on the website, which was effective to download a sound recording in Australia that was stored on a remote computer, there was an infringing act. Mr Cooper created and operated the website. He could have prevented these infringing acts, either by not establishing the link in the first place or, subsequently, by disabling or removing the link.

In the circumstances, it was not reasonably open to Mr Cooper to claim mere indifference to the use internet users made of the website. The findings at first instance as to the nature, the contents and structure of the website, which were not seriously contested, plainly supported the further finding that Mr Cooper deliberately designed the website to facilitate infringing downloading of sound recordings. Mr Cooper’s position was, in this respect, entirely different from that of the manufacturers and vendors of blank tapes, which was considered in Australian Tape Manufacturers.

Nature of the relationship between Mr Cooper and the Interntet user/operator of remote websites

Kenny found that there was a ‘relationship between [Cooper] and the remote website operators when he created the facility for them to put links on the website that he operated’, that that relationship, and the one with website users, was commercial in nature. Branson J only found a relationship with the users. Both emphasised the importance of the ‘commercial’ nature of the relationship, in reasoning reminiscent of some of the US courts.

Reasonable steps to prevent infringement

The final issue is whether Cooper took ‘reasonable steps to prevent’ infringements. This is the last of the factors courts have to consider. Both judges simply note that none were taken. Neither judge offers any thoughts as to what might be ‘reasonable steps’.

More striking, on this score, is what the judges have to say about the ISP. Both judges suggest that the ISP had the power to prevent, and failed to take reasonable steps to prevent the ifnringements. The evidence? The ISP could have withdrawn hosting. Yes, that’s right – because the ISP did not withdraw hosting, it was liable. Safe Harbour anyone? Not under current Australian law, where only Carriage Service providers get safe harbours from copyright infringement.


The difference bewteen the judgments, finally, I think, is illustrated in some of the concluding comments. Here’s Branson J:

‘Having taken into account the matters identified above, and the name of his website, I conclude that Mr Cooper infringed the Record Companies’ respective copyrights in sound recordings by in Australia authorizing internet users to do acts comprised in those copyrights, namely make copies of the sound recordings. I also conclude that Mr Cooper infringed the Record Companies’ respective copyright in sound recordings by authorizing operators of remote websites to communicate those sound recordings to the public in Australia.

Here’s Kenny J:

So far as internet users and remote website operators were concerned, the website was in substance an invitation to use the hyperlinks provided and to add new links in order that sound recordings could be downloaded from remote websites, and a principal purpose of the website was to enable infringing copies of the downloaded sound recordings to be made…. I agree with Branson J that Mr Cooper authorized the infringement in Australia of the record companies’ copyright in sound recording by authorizing internet users to make copies of sound recordings in which copyright subsisted, and operators of remote websites to communicate these sound recordings to the public.

Holistic analysis, anyone? Can you see why I prefer Kenny J’s views: who seeks to draw all these threads together, and view the situation holistically?

You see, here’s the thing. Authorisation, on precedents like Moorhouse and others, has to be, in the end, about whether, taking all the circumstances into account, the person has truly ‘sanctioned, countenanced or approved’ copyright infringement in circumstances where there is a sufficient connection between their acts, and the infringers, that they should be held liable for the infringer’s actions. the ‘Connection’ is ordinarily found in a combination of control over the infringer, knowledge of the infringements, and the relationship (whether of control, financial benefit, etc). Traditionally, and including if you go back to Moorhouse, the origins of this kind of liability do lie in the kind of principal-agent authorisation scenario, although obviously, over time, the law has developed beyond that original narrow context.

But liability under this ground isn’t – or shouldn’t be – wholly at large. In the end there is a question whether the person should be deemed ‘responsible’ for the other’s infringements. That’s where this kind of liability starts. Branson J’s judgment, in particular, I think, is so focused on the legislation that it loses sight of that origin. As a result, she comes out with comments that are deeply troubling, in that they seem to suggest very broad liability for any operator of a website, who has the power, in the end, to take any link down.
It’s easy to argue Cooper should have been liable, given all his conduct. But I think it is Kenny who nails the reasons: he did issue an open invitation and made available all the means and technology for infringement. That might be authorisation. A simple link is not. And the reasoning must draw the line between the two.

Final thoughts

I’ve given the general flavour of my thoughts at the beginning. THe most striking thing, in many respects, about these judgments – particularly that of Branson J – is that it seems relatively untroubled by the fact that on her reasoning, taken literally, many, many websites, and ISPs, could be liable for authorising copyright infringemetn -with little or no protection from the law. I think it’s sad that in this area – one of the few where common law developments are critical – that more effort wasn’t made to look beyond the borders of this little case, and craft a judgment that wouldn’t lead to quotes that make me, a blogger, slightly nervous about linking to infringing sites.

There’s a lot more I could say about these judgments. Some food for thought for the true copyright geeks:

  1. Branson J explicitly relies on the presence of section 112E to justify a broad view of authorisation. It wouldn’t be necessary, she says, unless there was some possibility that just providing facilities could lead to authorisation liability. What will that mean for some of those ‘avoiding doubt’ provisions we’ve just seen go into the Copyright Act?
  2. If an ISP can be liable in these circumstances, when is the government going to do something about online service provider liability under Australian law – so much stricter than US law where safe harbours are available to all online service providers?