I should perhaps start with saying that Proposals A and B are not miles apart. They both finally make it crystal clear that the European legislator desperately wants to implement ‘notice-and-staydown’ in the EU. Why then was there so much denial up until now? You can fill in the blanks for yourself and remember this next time academics are accused of misrepresenting the Commission’s proposals. According to both proposals, every ‘information society service provider that stores and provides access to the public to a significant amount of copyright protected works or other subject-matter uploaded by their users who do not hold the relevant rights in the content uploaded’ shall, in the absence of an agreement, ‘prevent the availability on their services of works or other subject-matter identified by rightholders’.
Who will bear the burden?
Online platforms will gain a new obligation ‘to prevent’ third party infringements, as long as they qualify as a) an ‘ISSP’, b) ‘storing and providing access to the public’ and their services carry c) a ‘significant amount’ of protected content (A:1;B:1a). This applies to all platforms regardless of whether they are established in the EU, or just provide relevant services to the EU (A/B:2a). Both proposals now include clarifications that storage or provision of public access are cumulative, and thus only storage-providing services, or only access-providing services won’t qualify. According to newly proposed recitals (A:38a; B:38b), this should exclude services ‘such as internet access providers, or providers of cloud services which are used by users to upload content for their individual use or to online marketplaces which are not used by users to access and consume digital content online but rather to provide access to works in the context of their main activity’. Is this clarification clear to you? Because it isn’t to me. Cloud providers and online market places both make copies. They have to. Online market places provide public access to offerings of their users. If offerings are tangible, you cannot enjoy them. But if they are intangible (e.g. files for 3D printing or a marketplace for pictures), you can easily consume them in the same way as, say, videos on YouTube. Cloud services don’t provide public access by default, but they can easily be turned into such public sharing, e.g. via Dropbox links. How is that any different from any cyber locker then? And where is the legislative anchor for ‘are not used by users to access and consume digital content online but rather to provide access to works in the context of their main activity’ anyway? Isn’t YouTube doing both of these things?
Furthermore, the recitals of both proposals (A:38a; B:38b) now clarify that ‘[n]or should it apply to providers of online services where the content is mainly uploaded by the rightsholders themselves or is authorised by them’. I am baffled by this carve-out. So, if YouTube attracts many original creators (as it does) can it absolve itself of the stay-down obligation? And if yes, how many are required? Does ‘mainly’ mean more than 50%? By the way, you realise that users who engage in mash-ups of content, whether permitted or not, are also right holders? As long as they meet our low thresholds of originality, they are also right holders, even if they create at the expense of infringing rights of other right holders. In other words, you could easily argue that many UGC-websites are stuffed with over 50% content that is uploaded by right holders themselves. I am really not sure about the intentions here.
‘[A] significant amount of content’ will need to be assessed ‘on a case-by-case basis and take account of a combination of elements, such as the total number of files of copyright protected content uploaded by the users of the services and the proportion of the protected content uploaded by the users in the overall amount of content available on the service.’ (A:38b;B:38c). Who will figure this out in practice? This is an invitation for fragmentation on the national level, even if you did support the proposal. First, you have no idea what is the standard. Does significant mean something like substantial in the Database Directive? If yes, we have a perfectly low threshold. Second, which videos count? How do you identify which videos uploaded are copyright-protected in general? If own-uploads by right holders don’t count, will you subtract these and then compare against overall non-copyright content? These are nice criteria in theory, but I am afraid they are not workable and you eventually end up with a ‘you-know-when-you-see-it’ test.
Nature of the obligation
Both proposals now make clearer that the legislator is attempting to establish a new stand-alone obligation to prevent third party infringement, independent of exclusive rights. It will be a copyright-related obligation, to be sure, but its infringement can hardly be seen as copyright infringement. Will the Enforcement Directive apply to it? We don’t know. Will the obligation be subject to the country of origin principle of Art 3 E-Commerce Directive (and thus moderated via establishment) or excluded as an IP issue? Again, we don’t know. In any case, the sanctioning of this obligation will be left to the Member States (A/B:4) which means additional fragmentation (this detail wasn’t even considered before). So we might end up with hefty administrative fines in one country and private claims in the other. And we will call that harmonisation. An interesting conflict might still arise once sanctions are imposed on platforms for not respecting Art 13. If they qualify for a safe harbour, they could argue that such sanctions are a form of ‘liability’ for third party content. And since the Directives are not meant to touch, I have no idea what the result would be.
Generally, the difference between the two proposals is that while Plan A ends here, B further tries, at the same time, to inflate the right to communication to the public by the already-known misrepresentation of the CJEU’s case-law (B:1) [fun-fact: you now have to also store information in order to qualify for communication to the public as a platform].
Another interesting thing is that Proposal B wants the ’Member states [to] provide that agreements concluded between information society service providers which store and give access to works and other protected subject matter uploaded by their users refererred (sic!) in paragraphs 1 and 1a above, and relevant rightholders, shall cover the liability of the users of the information society services, when they are not acting in a professional capacity, for acts falling within Article 3 of Directive 2001/29/EC’. ‘Covering liability’ is apparently a new concept of the EU copyright law. This wording targets the agreements of a) providers directly infringing rights by communication to the public [after the ‘restatement’] and b) providers not infringing in such a way. Is this meant to say that such agreements cannot cover acts of professional users of services who engage in communication to the public? Clearly, the concern is that once providers start paying they could absorb the entire licensing market for individual users as well. As I read this, and please correct me if I am wrong, the provider may only license on behalf of its non-professional users as a consequence. Of course, it can also license its own use, if there is any. But if this is the case, then isn’t the consequence that the agreements cannot prevent a provider continually being exposed to liability for professional content, unless they check their licensing first? If you yourself communicate because you are too actively intervening in other people’s communication, and cannot license for all your users, it means that for all non-licensable users – here professional users – you keep being exposed or at least have to take them down, despite the fact that you yourself have a licence. I know this is a bit of a mouthful. In other words, it appears to me that unless both a professional user and a communicating platform have a licence cumulatively, the video must go down. So a platform can be licensing itself for its professional users in vain, if they don’t license on their own as well. I could also imagine an alternative reading, namely that when a platform is licensing itself, automatically, it licenses also for non-professional users. However, in order to license professional users, another stipulation is necessary. Anyone?
Scope of the obligation
Both proposals stipulate that ‘[t]he measures shall be applied by the information society service providers at the request of rightholders to specific works and other subject-matter as identified by them. The measures […] shall be appropriate and proportionate, taking into account, among others, the nature of the services, the type of works or other protected subject-matter uploaded by the users of the services, the availability and costs of relevant technologies and their effectiveness in light of technological developments’ (A:1;B:1a). Accompanying recitals explain that ‘[t]he assessment of the appropriateness and proportionality of measures to be taken by the information society service providers should among other things take account of the type of content uploaded by their users, the state of the art of existing technologies per type of content and the size of the service. Where different categories of content are uploaded, such as music, text and audiovisual content, different measures may be appropriate and proportionate per type of content, including content recognition technologies’ (A:38c;B:38d).
As you can see, content recognition technologies now make it only to the recital. This is substantially softened language, to be sure, and there is clear reference to criteria to be considered, although they are not exhaustive. Given the completely open-ended nature of this provision, it is clear that it is an invitation to the CJEU and domestic courts to flesh out what are really the required measures. In this sense, the measures prescribed don’t appear any different from injunctions against intermediaries as legislated in Art 8(3) of the InfoSoc Directive. The important difference, of course, is that Art 13 is a market-entry requirement and thus applies to all qualifying providers, not only to a few defendants.
If this wording is passed as the law, I think there are, at least, two scenarios: (1) the CJEU will try to synchronise the scope of Art 13 obligations with what is possible under Art 8(3) InfoSoc. If this is the case, then the right holders might be not so happy to find out that stay-down might not work out the way they expect. Tommy Hilfiger C-494/15, which, in my view, is the last and leading CJEU reading on specific monitoring, limits preventive measures to ‘avoiding new infringements of the same nature by the same market-trader from taking place’ (para 34). German-style stay-down doesn’t pass this test because it does not care about who the infringer is. Let’s remind ourselves that the Estonian proposal now also adds reassurance for the E-Commerce Directive. (2) The second alternative is that the CJEU might read Art 13 as a political signal and the case-law might shift towards measures that are broader in scope than Tommy Hilfiger suggests. In any case, the outcome that is 100% certain is that obligatory preventive measures a la Art 13 will set a precedent and send a strong message outside the European Union.
Unlike in the Commission’s proposal, in this proposal there is stronger wording regarding the right holder’s duty to provide necessary input for the data of such preventive technologies (‘shall provide an information society service provider with the necessary data.’). Moreover, what remained is that, according to the Proposal, the Member States should facilitate ‘the cooperation between the information society service providers and rightholders through stakeholder dialogues to define best practices, such as the use of appropriate and proportionate […] technologies’. I cannot keep wondering why such dialogue should take place only between providers and right holders. Or are user-creators also meant as right holders? What about passive consumers or regulators? Such deals can be a hotbed for anti-competitive agreements and have an important impact on fundamental rights, so we should certainly demand more transparency.
What really put me off when trying to interpret the proposal is the formulation of the counter-notice. Although the wording stresses that Art 13-induced measures ‘shall be implemented by an information society service provider without prejudice to the freedom of expression and information of their users and the possibility for the users to benefit from an exception or limitation to copyright’, the important bit comes afterward. According to both Proposals, ‘[f]or that purpose the service provider shall put in place a complaint and redress mechanism that is available to users of the service in case of disputes over the implementation of the measures. Complaints submitted under this mechanism shall be processed by the relevant rightholders within a reasonable period of time. The rightholder shall duly justify its decision.’ In other words, as I read it, the right holders will not be consulted on counter-notices, but they will be the ones deciding on them. I really have a hard time understanding how this is compatible with freedom of expression. Copyright law carves out exceptions from the exclusive rights because certain socially relevant uses cannot be subject to the right holder’s reservation. For instance, how would society look if you couldn’t cite someone without his/her permission? According to the proposal, you might have an exception or be entitled to a public domain use, but when such use takes place on an online platform, and if a right holder disagrees with that, then you are released to his/her mercy.
It is hard to reach a conclusion since things are rapidly moving forward. At this week’s Julia Reda event I heard Mr Axel Voss saying that Art 13 is needed because there is no alternative to it. Yes, there is. It is called properly harmonised and standardised Notice and Action. All the claims about value gap or value transfer are based on the assumption that platforms are in a position to easily grab value from creative works. This would be true if the right holders had no credible threat to remove the content under notice and takedown, so they would have to de facto tolerate the use. But is this happening? I am not so sure. All the research I know (hat tip to CREATe) says that notified content gets removed very effectively and the problem is rather content that goes unnoticed. And yes, I agree that small rights holders might have a problem accessing tools used by big right holders. But again, this does not mean that the only solution is to mandate filters on everyone which only transfer the same problem onto small providers. So while big guys have their fight about who pays the bill, SMEs get hurt on either side.
The same preventive technologies are used today to detect and notify infringements. The governments can support development of cheap and high-quality technologies by intervening on the side of demand to help out the small right holders. Plus, if such solutions are open-sourced, they will not only be transparent but also widely available. CMOs and others can work together on perfecting them and offering them to their clients. So when people argue that Art 13 is the best way to develop effective technologies (e.g. Professor Giuseppe Mazziotti at Reda’s event), I have to disagree. I think that the incentives for development of such technologies are stronger, for a number of reasons, if the market is on both sides – right holders and providers – and not only on the side of (increasingly concentrated) providers.
To make sure you do not miss out on posts from the Kluwer Copyright Blog, please subscribe to the blog here.