The ADA is concerned that important amendments to Australia’s website blocking scheme which remove public interest protections and significantly increasing its risk of overreach and abuse are being pushed through without proper discussion, justification, or the support of the majority of stakeholders. We expand on these ideas in our submission to the Senate Standing Committees on Environment and Communications Inquiry on the Copyright Amendment (Online Infringement) Bill 2018.
Tuesday 20 November, 2018
Senate Standing Committees on Environment and Communications
PO Box 6100
Canberra ACT 2600
To whom it may concern
Copyright Amendment (Online Infringement) Bill 2018
The Australian Digital Alliance (ADA) welcomes the opportunity to provide comments to the Legislation Committee on the Copyright Amendment (Online Infringement) Bill.
The ADA is a non-profit coalition of public and private sector interests formed to provide an effective voice for the public interest perspective in copyright policy. It was founded in 1999 to unite those who seek copyright laws that both provide reasonable incentives for creators and support the wider public interest in access to knowledge, advancement of learning, innovation and culture. ADA members include universities, schools, disability groups, libraries, archives, galleries, museums, technology companies and individuals.
Website blocking is a drastic remedy and a blunt tool which has significant implications for free speech, innovation and the functioning of Australia’s internet. As such we are concerned that important amendments to the scheme are being pushed through without proper discussion, justification, or the support of the majority of stakeholders. These amendments removed public interest protections that were touted as central to the original scheme, significantly increasing its risk of overreach and abuse.
The Bill makes substantial changes to established, complex law – law which according to the testimony of almost all stakeholders is currently working effectively. The need for these changes is questionable, as both stakeholders and independent parties have stated that the established process is fair and works well for all parties.
The proposed amendments significantly expand the scope of the scheme, while at the same time reducing judicial oversight and increasing its remedies. This is a highly concerning combination of amendments for a scheme which in 2015 was explicitly sold as setting “an intentionally high threshold test for satisfaction by the Court” to limit the scheme only to “those online locations which flagrantly disregard the rights of copyright owners”. It appears that the protections for the public interest included in the scheme are being eroded without supporting evidence and without the opportunity for proper public debate or an understanding of their likely effects.
In particular, we recommend that the Bill be amended to:
- Remove the proposed “primary effect” test;
- Restore judicial oversight to the scheme by removing the process by which ISPs / search engines are forced to negotiate with rights holders on further sites to block;
- Reverse the inefficient and ineffective decision to require search engines to block sites already inaccessible to Australian users, or to include search engines only if they are granted the same rights as ISPs under the copyright safe harbours.
We thank you for the opportunity to comment. Our principal contact with respect to this matter is our Executive Officer, Jessica Coates, who can be reached at firstname.lastname@example.org or on 02 6262 1118.
Website blocking should reserved for exceptional cases
Blocking access to a whole website is an extreme step, that should only be taken in rare cases to deal with large scale piracy sites. Blocking access to whole sites has strong and potentially negative effects on free speech, public discourse and the rights of individual Australians. Furthermore, it potentially serves a killer blow to any legitimate website incorrectly caught in the scheme, due to the material revenue, brand and reputational losses that can be caused by any significant downtime. As such the scheme has great potential for abuse, whether intentional or through error, and its remedies should be applied only cautiously and with high standards of proof. The scheme certainly should not be expanded to cover an increasingly large number of websites, which to date have remained unidentified by those seeking the amendments.
Little stakeholder support for change
The ADA does not believe the case has been made for the expansion of Australia’s website blocking scheme at this early stage. We note that the evidence provided in submissions to the government’s Review of the Copyright Online Infringement Amendment earlier this year, from both rights holder advocates and other stakeholders, is that the existing scheme is working. For example, the Australian Copyright Council states that “the site-blocking scheme has had a material and effective impact on the level of access to websites facilitating online copyright infringement.” Similarly, PPCA states that “it is an important mechanism for providing online copyright protection and assists in reducing the accessibility of overseas infringing websites in Australia.” A number of recent studies also indicate that infringing activity in Australia has dropped since the legislation was introduced.
In contrast to requesting changes, the scheme has been so effective that a large number of rights holder representatives (including Music Rights Australia, the Phonographic Performance Company of Australia, Australian Recording Industry Association, and APRA AMCOS) specifically requested in their submissions that no changes be made at this time, and instead recommend that “the Government conduct a formal review of all aspects of copyright enforcement within eighteen months [to] ensure s115A is an effective and efficient mechanism in the context of the overall rights protection environment.” While the ADA does not believe a formal review of all aspects of copyright enforcement is necessarily warranted, it is clear from the majority of rights holders’ representatives’ submissions that changes to the website blocking scheme at this time are not required.
The principal criticism put forward by those rights holders seeking immediate amendment (Australian Film and TV bodies) is the length and cost of obtaining orders.
In fact, only three submissions to the government’s review requested changes akin to those proposed in the Bill. Foxtel and the Sports Coalition requested the lowering or removal of the “primary purpose” requirement, while Foxtel and Village Roadshow requested that the scheme be extended to search engines. None of the public submissions appear to have proposed that the scheme be broadened to allow blocking through private agreement.
“Primary effect” substantially lowers the threshold for blocking
Our principal objection is to the proposal to extend the scope of the Bill to allow the blockage not just of sites that have the “primary purpose” of infringement, but also those that merely have the “primary effect” of infringement. This lowering of the bar directly undermines the “intentionally high threshold test” of the original legislation, opening the scheme up to the potential for significant unintended consequences. This substantially broadens the scope of the scheme, and opens up the possibility that it could capture not only malicious piracy sites, but also sites which have been co-opted by piracy communities, and sites that have significant legitimate purposes but are nevertheless infringing under Australian law.
Lowering the threshold requirements of the scheme to allow blockage of websites with only a “primary effect” of infringement holds the danger of allowing websites to be blocked that have been co-opted, without the intent or possibly even knowledge of their owners, for infringing purposes. Because of the increasing prevalence of site blocking schemes around the world, pirates are increasingly looking to use innocent services to host infringing content. It has even been reported that pirates have used Google Maps to host pirated content after Google strengthened its enforcement on Google Drive. Other smaller sites have also been targeted by pirates, such as the website change.org. It is not hard to see how smaller websites with insufficient resources or expertise could quickly be overrun by nefarious actors.
It would also create potential difficulties for fledgling user generated content platforms, such as a new social media platform, which is unlikely to have the resources to invest in sophisticated copyright removal technology at an early stage. This creates a high risk of websites which clearly do not have a primary purpose of facilitating infringing activities, but which may be used by bad third party actors for such activities – at least for a time – being the subject of site blocking orders.
Furthermore, the very broad “primary effect” language also has the potential to capture a large number of legitimate and commonly used websites. The Explanatory Memorandum of the Bill explicitly states that the “primary purpose or primary effect” test will exclude online locations that are “mainly operated for a legitimate purpose, but may contain a small percentage of infringing content.” However, this ignores the reality that, due to the lack of general exceptions in our Act, a large number of innocent and everyday activities currently infringe copyright in Australia, including generating memes, auto-translation, cloud storage, caching and indexing, and sharing screenshots of websites. This has been identified as a major failing of Australia’s copyright system by a number of reviews, including most recently the Australian Law Reform Commissions’ (ALRC) Copyright and the Digital Economy report (2014), and the Productivity Commission’s (PC) Inquiry Report on Australia’s Intellectual Property Arrangements (2016).
This means that sites which are perfectly legal in other jurisdictions with more internet-appropriate copyright laws, such as the United States, infringe copyright in Australia. This includes a large range of innocent and commonly used websites and apps such as:
- Pinterest – as no exception allows the sharing of screenshots and images for informational purposes;
- auto-translation services – as no exception allows translations for general purposes, meaning you would have to have a licence or prove that the translation was for research and study, reporting the news, criticism and review, parody and satire, or legal advice, for it to be legal;
- popular meme creation sites such as Meme Generator, Meme Maker and iMeme – while in some memes created by these sites may be permitted as a fair dealing for parody and satire, many memes are not necessarily humorous, and it is not clear that the sites themselves would be protected by this defence; and
- cloud storage providers – with the Federal Court having confirmed a commercial cloud storage service can infringe copyright even where the activities it enables would be legal for its clients.
While these services clearly don’t have a “primary purpose” of infringement, it is easy to see how they could be found to have a “primary effect” of infringement, as the majority, if not all, of the activities they enable are infringing under Australian law. They certainly do not fall within the narrow carve out for sites that “contain a small percentage of infringing content.”
Similarly, although the scheme allows courts to take the flagrancy of infringement into account, and whether the site operator demonstrates a disregard for copyright, this provides little protection for such sites. This is because they arguably are flagrant as to their facilitation of infringement in Australia, and demonstrate a disregard for local copyright law, in that they allow Australians to make use of their services without warning, geoblocking or otherwise seeking to limit infringements in this country. Other factors such as how proportionate a blockage would be are similarly open to interpretation – for example, an Australian court recently found that blocking could be proportionate even if a significant portion of the material being shared was non-infringing.
Piracy v infringement
If the scheme were aimed at sites with a primary effect of piracy, we would not be so concerned. Piracy is an act which involves circulation of whole copies of copyright works with the intent of replacing sales. It clearly harms copyright owners, and sites which primarily facilitate piracy are justifiably the target of legal action. The Copyright Act does not currently include the concept of piracy; however it does have a close equivalent in a criminal offence set out by s132AC, which targets “commercial-scale infringement prejudicing the copyright owner.”
However, these concepts – that the activity in question be commercial scale, or that it be prejudicial to the copyright owners – are not incorporated into Australia’s website blocking scheme. Instead, sites can be blocked on the basis of facilitating mere infringement, a far broader concept that, as discussed above, incorporates many innocent and common activities. As currently drafted the Bill does not require evidence that a site is linked to piracy (as opposed to mere innocent or incidental infringement), or that it has caused harm or prejudice to copyright owners at all. Yet these should arguably be a basic requirement for any site to be blocked – that it be shown to actually be causing harm, or at least prejudice, to the copyright owner.
Creep is already happening
We recognise that it is unlikely that any rights holders will attempt to bring action to block Pinterest tomorrow. However, we are very concerned about the “creep” in the scheme that these changes represent, that the gradual expansion of its boundaries with little justification undermines the “intentionally high bar” of the original scheme, opens the door to abuse, and steps outside the expectations of the Australian public.
Over the last two years four cases have been brought under the scheme by representatives of the Australian film, music and subscription television sectors which, as expected, target traditional piracy sites such as Pirate Bay, KickassTorrents and LimeTorrents. Two more recent decisions have expanded its focus to sites that stream illegal content via “smart boxes” – again, fairly much as expected, though it is notable that in one of these cases the court agreed to block a site where only some of the content streamed was illegal. But the most recent case, still before the court, includes in its targets an entirely new type of service – so called “fansubs” sites, which make available amateur subtitle files of films in alternate languages. These files are used in conjunction with pirated or legitimate copies of films to make them accessible to those for whom the film’s language is not their native tongue, or who are deaf.
International film industry groups criticise fansubs sites as harmful to the market because they increase the utility of pirated material and undermine region-based business models. And there is no doubt that in many cases their activities are an infringement of copyright in Australia, as film scripts are protected as separate copyright works, of which the subtitles would be an unauthorised adaptation. However, it is questionable whether such sites represent the “worst of the worst” that were declared as the intended target of the initial website blocking regime. Certainly, they are unlikely to fall within the general public’s definition of a piracy site. The do not distribute full or even partial copies of films, and can be used in conjunction with legitimate content as readily as illegitimate. They rae often the only source of subtitles for a film beyond a small number of extremely common languages. Even the trial judge for the case, Justice Nicholas, reportedly questioned the “creep” from earlier site blocking cases and whether it would survive the “primary purpose” test.
Once fan subtitling sites are blocked, it is arguably an easy step to meme websites, for example, as memes are often turned into unauthorised merchandise for which copyright owners receive no compensation. Indeed, Australian-based design marketplace Redbubble has already been sued by both Pokemon and Hells Angels for unauthorised merchandise appearing on their website.
Similarly, VPNs which provide important privacy and data protection services could be targeted for providing unauthorised access to the international catalogues of services such as Netflix. The issue of VPNs was raised as a specific concern by the Senate Legal and Constitutional Affairs Legislation Committee when it examined the 2015 Bill, and on its recommendation amendments were added to the Explanatory Memorandum (EM) to note that VPNs were not intended to be captured by the Bill. However, these assurances were based very much on the “primary purpose” test and are undermined by the removal of this requirement. To quote the EM:
The [primary purpose] test is also not intended to capture Virtual Private Networks (VPNs) that are promoted and used for legitimate purposes, or merely used to access legitimate copyright material distributed in a foreign geographic market. VPNs have a wide range of legitimate purposes and have no oversight, influence or control over their customers’ activities.
Based on the argument above, it is easy to see that a “primary effect” test may well capture VPNs. The fact that the VPNs “are promoted and used for legitimate purposes” has relevance under a “primary purpose” test, but not under a “primary effect” test, particularly if an argument were put forward that the main use of VPNs was to hide piracy activities or gain unauthorised access to overseas material. Yet there is no corresponding language in the new EM to clarify the intentions under the new language towards VPNs.
The point is that lowering the bar for website blocking without strong and clear justification, and without including sufficient safeguards to ensure that only true piracy sites are captured, is extremely problematic. We emphasise again that website blocking is an extreme and draconian measure that should only be exercised in rare cases of clear malicious intent. We are not aware of any evidence that supports the need for this amendment, especially as rights holders have almost universally stated that the scheme in its current form is successful in blocking major infringing sites. We are not aware of any applications for injunctions which have failed based on the current test. It is not clear why some feel that the requirement to prove “primary intent” cannot be satisfied by reference to the behaviour of the website managers with regards to infringing activities eg whether they respond quickly to takedown requests, whether they promote or encourage the infringing activities, whether they have taken steps to block repeat infringers.
Blocking through private agreement
The ADA also strongly opposes the proposal to establish a process that would require carriage service providers and search engine providers to negotiate with rights holders to agree on expansion of court order blocks (ss115A(2B)(a)(ii) and (b)(ii)). This reduces judicial oversight of the scheme, pushing decision-making as to the legality or otherwise of websites into the hands of rights holders. Based on past experiences of negotiation between rights holders and CSPs, it is likely that the resulting process would be costly, lengthy, and problematic, and arguably would be less efficient than a Court order, particularly considering the established abridged process for expanding blocks. The current successful processes for the scheme having been carefully negotiated at length over previous court cases. The proposed changes disrupt these well established processes, requiring new negotiations and increasing costs and delays for all parties.
Due to the serious and extreme nature of the potential impact of the scheme, it is essential that close judicial oversight be maintained to ensure that websites are not erroneously blocked or blocked without clear evidence of malicious infringement. Any amendments which imply that rights holders will be given much greater latitude in determining which websites to block, or that ISPs may be required to monitor the communications of their clients for copyright purposes, should be avoided, to ensure the integrity not just of the scheme but of the Australian legal system itself.
The Federal Court has itself indicated that follow on orders are entirely a matter for it. In the judgement for Foxtel Management Pty Ltd v TPG Internet Pty Ltd & Ors NSD 241 of 2016, Justice Nicholas stated:
In the case of a variation to an existing injunction, the Court may be willing to act on very little in the way of further evidence. Nevertheless, whether or not there should be such a variation remains a matter for the Court.
Justice Nicholas’ comments also suggest that follow on orders could be sought cheaply and efficiently from the Court. This was confirmed in the case of Roadshow Films Pty Limited & ors v Telstra Limited & ors NSD1925 of 2017. In that case Justice Nicholas made an order that provided a very simple and effective method for the applicants to get site blocking orders in respect of websites that begin to operate from a different Domain Name, IP Address or URL after the date of the initial order. The ADA endorses that approach as one that is to be preferred to private negotiations between commercial entities.
We urge the government to maintain full judicial oversight of any orders made under the scheme, to ensure that they meet the standards of evidence required by a democratic legal system.
Extension to “online search engine providers”
Finally, we believe that the extension of the scheme to online search engine providers is inefficient, ineffective and contradicts recent government policy.
Firstly, we note that the government recently, through the Copyright Amendment (Service Providers) Act 2017, made a clear policy decision that non-carriage service providers, including search engines, were legally different from carriage service providers, and do not have the same legal rights or responsibilities. The amendments contained in the Bill seem to contradict that decision. If the scheme is to be extended to search engines, it should only be done on the basis that they are also granted protection under the copyright safe harbours.
Notwithstanding that contradiction, it is unclear how adding search engines to the list of service providers covered by the website blocking scheme would change access to the sites in Australia when they are already blocked at the ISP level. Material on these sites will already be unavailable to users searching from Australia. As such, this proposed extension has no clear purpose or effect. Its one major impact will be that it will materially increase the costs of the scheme through “double dipping” – increasing the number of organisations rights holders and the Federal Court must deal with and the number of organisations that are required to take action – all to achieve the same result.
It is telling that of the many countries that have enacted site blocking legislation, including countries throughout the EU, the UK and South Korea, not a single country in the world has extended its site blocking schemes to search engines. There is no good reason why Australia should be the first.