Filtering measures, freedom of expression, and user rights
Prior to the finalisation of the transposition proposal, the government obtained (obligatory) legislative advice from the Council of State. This advice had raised the issue of Article 17(4)(b) measures in relation to freedom of expression, albeit without coming to any conclusions. In its response (chapter 6 of the explanatory memorandum), the government summarises the relevant EU Charter provisions and CJEU case-law on filtering injunctions and balancing fundamental rights. In the penultimate paragraph, the government acknowledges that filters can match content but cannot understand context, and thus cannot detect when matched content is lawful because an exception or limitation applies. However, it maintains that users’ free expression rights are adequately protected, because they can use the complaint and redress mechanism prescribed by Article 17(9):
It should be borne in mind that under certain conditions and circumstances using copyright- and neighbouring rights-protected works is permitted, for example on the basis of the right to quote or the parody exception. Assessing these situations through the intended filtering measures is in any case not yet possible according to the current state of the art. In order to reduce the risk that the freedom of expression and of information of users who could invoke such a restriction will be unfairly affected, the Directive and, like it, the legislative proposal contain a specific safeguard. This safeguard is also in line with the findings of the Court of Justice of the European Union in the aforementioned Sabam/Netlog NV judgment. The safeguard stipulates that the provider of the online content sharing service must provide for a complaint and redress mechanism. Users should be able to have recourse to it in the event that access to and use of protected services is restricted even if it is subject to a restriction. The aim of this mechanism is to ensure that legal content is not blocked or removed as a result of the inadequate application of the envisaged filtering measures, quickly, inexpensively and through human review. The article-by-article part of the explanatory memorandum outlines how this mechanism works. In brief, a user who believes that material uploaded by him has been unfairly blocked may report it and the rightholder will have the opportunity to respond to such a report. If such a response is not received or is not received in time, the content is simply made available. There is also a means of redress available to both rightholders and users in the event of an unwelcome decision. For the sake of completeness, it is mentioned here that the possibility of submitting a dispute to the public courts, as enshrined in Article 17 of the Constitution, is of course not affected. The complaints and appeals mechanism is an important link in the directive and the bill to achieve an appropriate balance between the conflicting fundamental rights.
The notion that user rights in relation to the application of filtering technology can be adequately protected using only counter-notice seems at odds with the balance that the EU legislator – in particular the European Parliament – sought to inject into Article 17. There is considerable empirical evidence to suggest that users are unlikely to counter-claim, regardless of whether they have plausible grounds to do so – with chilling effects disproportionately impacting women, and users mentioning fear, complexity and even uselessness as reasons not to contest a block or takedown. Moreover, counter-claim does not protect freedom of access, because users who are not the blocked uploader cannot counter-claim. While the government explicitly recognises that current filtering technology is not able to judge context, and thus cannot adjudicate the applicability of exceptions & limitations, it does not consider alternative conclusions, such as that it might not always be proportional, or compatible with Article 17(7), to require the use of such technology as a condition for invoking the 17(4) safe harbour. Neither does it consider alternative approaches, such as moving the ‘counter-notice’ phase forward by enabling users to assert the applicability of an exception or limitation during the upload phase (at least in those circumstances where file-matching could be done quickly enough to make this workable).
The discussion on filtering and freedom of expression in the explanatory memorandum could be considered a loyal attempt to defend the basic validity of Article 17. However, the transposition proposal contains other actual choices in relation to user rights that are not required by the directive, and indeed may not be compatible with it. Thus, the operative provisions do not contain a full transposition of article 17(7), which prohibits the prevention of the availability of works that do not infringe copyright or neighbouring rights, including where such works are covered by an exception or limitation. Similarly, the operative provisions do not transpose Article 17(8), to the effect that the application of the provisions may not lead to a general monitoring obligation (this is only added in the explanatory memorandum).
Other issues with the user rights provisions are the result of lacunae in Article 17 itself, which the government has either chosen not to address or felt unable to address. For example, Article 17(9) does not answer the question of who is liable for decisions taken in the context of the user complaint and redress mechanism: if a user objects to their upload being blocked and invokes an exception, the OCSSP may honour the objection and reinstate the upload – but is the OCSSP liable if a court later determines that the user could not invoke the exception? Conversely, if the OCSSP sides with the rightholder but is later overruled by the court or an out-of-court dispute-resolution body, are either the OCSSP or the rightholder liable for any damage suffered by the user?
Article 17(9) ought logically to have prescribed a safe harbour for good-faith decisions taken by OCSSPs in the course of complaint and redress proceedings. Arguably, it provides enough room for the Dutch transposition to include one, either by amending the proposal or through the order in council envisaged by Article 29c(7). In fact, the discussion in the explanatory memorandum on filtering and fundamental rights discussed above ends with a specific reference to the possibility of an order in council, describing it as a way in which additional safeguards for the freedom of expression might be provided if the balance tilts too far in the other direction.
For now, the current lack of clarity probably incentivises platforms to err on the side of caution (i.e. rightholders) when it comes to adjudicating disputes. Conversely, the lack of sanctions for (obviously) incorrect copyright claims incentivises rightholders to push back on users’ claims to invoke exceptions. Further, Article 29c (6) seems to require OCSSPs to ensure that rightholders and users have access to an impartial dispute-resolution committee, while Article 17(9) suggests that the dispute-resolution procedure is the responsibility of the member state. It is questionable whether a dispute-resolution committee installed by the OCSSP can be considered ‘impartial’, and here too the question of who is liable for carrying out dispute-resolution decisions is not addressed.
Taken together, these issues suggest that the balance intended to be ensured by Article 17 is not yet fully implemented in the transposition proposal.
The Dutch transposition proposal is now at the start of its parliamentary passage. The lower house will now exchange written comments with the government, after which the proposal will be debated – and potentially amended – in plenary session. The proposal will then move on to the upper chamber, which follows broadly the same procedural steps but has no right of amendment. On current timelines, timely transposition is still very much possible.
When it comes to transposing the press publishers’ rights and online content-sharing services, the government has a tough hand to play. Working from a difficult text that it didn’t support, and hemmed in by detailed provisions and polarised stakeholder input, the government has presented a proposal that is carefully calibrated to minimise risk. On balance, the proposal’s lack of original interpretative guidance on scope and definitions is probably a good thing, since any national guidance would have limited standing. Nonetheless, there are a number of issues on which amendments are worth considering, or indeed necessary in order to ensure correct transposition. For example, the transposition is more restrictive than the directive prescribes in relation to “short extracts” of press publications and the proportionality of measures taken by online content-sharing service providers. Moreover, a more creative and proactive approach would be both warranted and possible in some areas, such as user rights under Article 17 – which arguably should not be left to the market, to the courts or to secondary legislation that has yet to be conceived.
It remains to be seen how active an interest parliament takes in the transposition. On the one hand, there may be more pressing priorities, particularly given the limited room for national manoeuvre in the transposition phase. On the other hand, the directive did gain a certain notoriety during the Brussels negotiations, in particular on the issue of “upload filters”, and consumer digital rights’ groups are already taking an active stance on the transposition. While the government may be hoping that the other stakeholders cancel each other out, different parliamentary groups with half an eye on parliamentary elections due in 2021 might yet engineer some political fireworks.
 French Ministry of Culture, ‘Towards more effectiveness of copyright law on online content sharing platforms: overview of content recognition tools and possible ways forward’, Mission Report (January 2020), pp. 90-96; J. Penney, ‘Privacy and Legal Automation: The DMCA as a Case Study’ (September 2019), Stanford Technology Law Review, Vol. 22, No. 1, 412, https://ssrn.com/abstract=3504247; J.M. Urban et al., ‘Notice and Takedown in Everyday Practice’ (2017), UC Berkeley Public Law Research Paper No. 2755628, http://ssrn.com/abstract=2755628. See further references in this blog post by Daphne Keller.