Photo by Gerd Altmann via Pixabay


The long-awaited European Commission Guidance on art. 17 of the Copyright in the Digital Single Market Directive (CDSMD) was finally published on 4 June 2021, a mere working day before the transposition deadline of 7 June.  Since then, there have already been several pieces published weighing in on the Guidance’s merits and shortcomings (mostly the latter), namely from civil society outfits like COMMUNIA, EDRi and EFF, as well as in this very blog in two thoughtful comments by Reda and Keller, and Jutte and Priora (see also here).

The Guidance is adopted as a Commission Communication (COM(2021) 288 final), meaning that it is not legally binding. The Guidance is a 27-page document that is divided into seven sections: Introduction (I); a specific copyright authorization and liability regime (II); Service providers covered (III); art.  17(1) and (2) authorizations (IV); art. 17(4) specific liability mechanisms (V); safeguards for legitimate uses of content and complaint and redress mechanisms (VI); and transparency and information obligations (VII).

As a general remark, the Guidance is a mixed bag. As would be expected, it provides useful suggestions on certain points and falls short on others. But it is to a significant extent a document that denotes misgivings and hesitancy from its authors, who sometimes go to considerable lengths to avoid offering concrete guidance. This is most clear from the frequent recourse to the need for a case-by-case assessment (ten plus times) on key issues. As a result, Member States are often told between the lines to let the courts – and ultimately the CJEU – decide.

This is perhaps to be expected from a document published in the shadow of the action for annulment in C-401/19 (examined here), and which recognizes that it might need revising in light of that judgement. Nevertheless, the Guidance will at least be partially unsatisfactory for most of the national legislators that have waited for it until the last minute.

Most of the early commentary thus far focuses on the potential challenges arising from the interpretation of the preventive measures in art. 17(4)(b) and (c). Relatively little attention has been paid to the remainder of Guidance. This post focuses on some of these less examined aspects, namely the specific copyright authorisation regime, the different types of authorisations in art. 17(1) and (2), and the obligation to demonstrate best efforts to obtain an authorisation in art. 17(4)(a). A good part of the Guidance restates the legal regime of art. 17 CDSMD and supporting recitals. Since the provision has been extensively covered on this blog, this post focuses on the actual guidance provided in addition to what is stated in the law.


A Specific Copyright Authorisation and Liability Regime

The first major element of the Guidance is the unequivocal statement that art. 17 “creates a legal basis” for rightholders to authorise the use of their works when uploaded by users. Art. 17 is therefore viewed as more than a “clarification” of the existing law, as it changes the pre-existing legal framework, namely vis-à-vis art. 3 InfoSoc Directive.

The point is clear from the statement that art. 17 is lex specialis to arts. 3 InfoSoc Directive and 14 e-Commerce Directive. At the same time, it is stated that art. 17 does not introduce a new right in EU copyright law (p.3).

With these statements, the Guidance takes a position on that has been much debated in scholarship, in particular as potential legal qualifications of art. 17, as first laid out in this paper by Husovec and myself. A qualification of art. 17 as lex specialis is equivalent to what we call a special right (Option C). This means that the Commission rejects the potential options advanced therein.

On the one hand, it rejects that the right in art. 17 is fully encompassed in the minimum international standard (Option A) or in the EU standard of art. 3 InfoSoc Directive, as interpreted  by the CJEU (Option B). In other words, the right in art. 17 CDSMD is different from the right in art. 3 InfoSoc Directive. On the other, hand, it also rejects the option that art. 17 is a wholly new or sui generis right in the acquis (our Option D).

To be sure, the upcoming ruling in Peterson v YouTube (joined cases C-682/18 and 683/18) will have something to say about this (see here on the AG Opinion). Still, the Guidance’s rejection of the view that art. 17 is fully encompassed and follows the same regime as art. 3 InfoSoc Directive is clear.

An important consequence of the lex specialis nature of art. 17 is that the provision requires specific implementation into national law. That is to say, it is not possible for Member States to rely on their previous implementation of art. 3 InfoSoc Directive.

In this regard, the Guidance highlights a key aspect in the legal regime, which bears on the national transposition. That aspect relates to the fact that online content sharing service providers (OCSSPs) provide access to content originating from user uploads. The consequence is that national laws must explicitly introduce “specific conditions” that reflect this composite nature of the act of communication.

Another key statement in the Guidance is on the legal nature of user rights (or mandatory exceptions or limitations) and safeguards in arts. 17(7), (8) and (9). These are qualified as “obligations of result”, meaning that they must prevail in case of conflict with preventive measures in 17(4) (p.3). As noted by previous commentators (here and here), it is arguable that the Guidance failed to fully take its own qualification under consideration, especially with the introduction of a loophole on ex-ante safeguards in the form of “earmarked” content.


Art. 17(1) and (2): authorisations

Building on the qualifications above, section IV of the Guidance discusses the concept and scope of the authorisations mentioned in arts. 17(1) and (2). The following examines each in turn.


The authorisation in art. 17(1)

Art. 17(1) does not define “authorisation”, which is viewed as an open-ended concept. Member States can therefore provide for different authorisation models. An interesting clarification in this respect is that the concept of authorisation does not require a remunerative counter performance. Indeed, it is also acceptable that an authorisation is given for free (e.g., as with Creative Commons licenses) for “data or promotional activities”.

A more important note in the Guidance is that acts of communication/making available to the public in art. 17(1) will include any necessary reproductions for those purposes. The Commission therefore takes the view that there is a functional merger between reproduction and communication to the public. As such, no separate permission required for reproductions in this context.

It is also stated that it should be possible for rightholders to not grant an authorisation to online content-sharing service providers (OCSSPs) for acts of communication to the public under art. 17 (see also Recital 61, making reference to contractual freedom).

From these statements it results that certain authorisation models are possible for the right in art. 17(1). This is the case, first, of individual and voluntary forms of collective licensing, provided the requirements of the Collective Rights Management Directive are met.

Furthermore, at least in the context of individual agreements, the Guidance encourages Member States to use, establish or maintain voluntary mechanisms to facilitate agreements between rightsholders and OCSSPs (p.6, perhaps nodding at mechanisms similar to those envisaged in art. 13 CDSMD).

The Guidance also admits that this collective licensing with extended effect (CLEE) is also a lawful model of authorisation, provided it is compliant with the requirements of art. 12 CDSMD, and that art. 12 is actually transposed into national law. The text of the Guidance on this point betrays the EC’s discomfort with the obvious conclusion that CLEE must be an acceptable authorisation model.

What is not stated in the Guidance but appears to be the conclusion from the above is that in principle mandatory collective licensing and statutory licensing are not viable authorisation models, since they are not voluntary for rightsholders (or at the very least do not include an opt-out safeguard).

Perhaps the reason why this is not stated explicitly is because the conclusion is not uncontroversial. In fact, if it is accepted that the right in art. 17(1) is lex specialis, as proposed in the Guidance, then this theoretically opens the door for non-voluntary and hybrid authorisation models, within certain limits, namely the effet utile of the provision (for a detailed analysis, see here).

The authorisation in art. 17(2)                                                                                                           

The objective of art. 17(2) is legal certainty for as many uploading users as possible. The provision states that authorisations obtained by OCSSPs shall also cover acts carried out by users of the services. The scope of the authorisation for these users extends to all their uploading acts that are: 1) for a non-commercial purpose; or 2) generating non-significant revenues.

In defining these concepts, Member States are told not to pre-set quantitative thresholds for ‘significant revenues’. This means that these concepts must be defined pursuant to a case-by-case analysis.

One important clarification is that if the user has explicit authorisation to upload content, there is no need for additional authorisation to the OCSSP (see Recital 69). This subsequent availability is considered to be within the scope of the initial authorisation, and not a presumption in favor of OCSSPs.

However, the Guidance adds that it is up to rightsholders to expressly exclude subsequent uploading from the authorisation to end-users. To do so, the EC encourages sharing of information to prevent blocking of legitimate uses. Still, no real specifics are provided here and the proverbial hot potato is tossed to Member States, OCSSPs and users.


Best efforts to obtain an authorisation

Section V.1 of the Guidance addresses the specific liability mechanisms on art. 17(4) CDSM. These mechanisms encompass a series of “best efforts” obligations to: (a) obtain an authorisation; (b) ensure unavailability of specific protected content; and (c) put in place notice and take down and notice and stay down mechanisms.

In framing art. 17(4), the Guidance clarifies three important aspects. First, “best efforts” is an autonomous concept of EU law. Second, these “best efforts” obligations are hierarchically inferior – and therefore must bend the metaphorical knee to – to the obligations of result in arts. 17(7), (8) and (9). Third, the proportionality principle in 17(5) should “be considered” when assessing all best efforts obligations in 17(4). This latter consideration is particularly important to protect OCSSPs’ freedom to conduct a business in art. 16 Charter.

On this basis, the Guidance predictably states that the obligation of “best efforts” to obtain authorisation in 17(4)(a) requires case-by-case analysis of actions of OCSSPs to seek out and/or engage with rightholders.

More concretely, it adds as a minimum threshold of that obligation that OCSSPs engage proactively with rights holders that can be easily identified and located, notably those with broad catalogues (e.g., CMOs). On the flipside, OCSSPs should not be expected to proactively seek out rightsholders “not easily identifiable by any reasonable standard”. (On how the structure of art. 17 favors big rights holders, see here.) To facilitate this, the Guidance suggests timidly that Member States may encourage development of registries of rightholders to be consulted by OCSSPs, in compliance with data protection rules (p. 9).

The Guidance further notes that the obligations under art. 17(4)(a) should be asymmetric depending on the OCSSP in question. In simplified terms, larger OCSSPs (“with a big audience in several or all Member States”) must seek out “high number of rightholders”, at least for the content that is predominant on their services. OCSSPs that are smaller, i.e., with a limited or national audience, must seek at least CMOs and “possibly a few other easily identifiable rightholders.” Those small and new OCSSPs covered by art. 17(6) must also seek out licenses, although the extent of their obligations should be assessed on a case-by-case basis with consideration for the principle of proportionality. In any case, smaller OCSSPs must provide contact details to potential licensees and engage with these if contacted by them.

Against this basic set-up, the Guidance states that refusals to negotiate or unreasonable refusals to license by rightsholders result in OCSSPs meeting their obligation to seek authorisation under art. 17(4)(a). However, and herein lies the rub, it is up to OCSSPs to demonstrate this, which might be challenging in practice, at least in the context of individual licensing deals.

A relevant wrinkle is suggested for cases where a type of content is present but not common or prevalent on a certain platform. Here, OCSSPs do not need to proactively seek out licenses but must still “engage” with rightsholders that offer them.

This point is trickier than it seems. At first glance, it is legally sound, since art. 17(1) leads to direct liability for protected content hosted by OCSSPs. Therefore, where the content is identified, it is susceptible of licensing. But if content is less common or prevalent on a service, then an OCSSP might have no interest in hosting it, and could therefore legitimately choose to take it down (e.g., under art. 17(4)(c)). However, the structure of the provision means that in order to benefit from a liability exemption, OCSSPs must demonstrates best efforts under 17(4)(a).

What the Guidance is saying then is that OCSSP must make best efforts to license content in which they might not be at all interested. Seen in this light, this is quite a far-reaching statement, perhaps inconsistent with the proportionality provision in art. 17(5) CDSMD and providers’ freedom to conduct a business in art. 16 of the Charter.

Reading the Guidance, it is unclear whether OCSSPs they have an obligation to license the less common or prevalent content under the same terms as the content that is core to their platform. If yes, then the only difference in this scenario is that OCSSPs are not required to proactively seek authorisation. On the other hand, and as potential mitigation on this obligation, it could be argued that this content is of less importance and value to the OCSSP. Therefore, the obligation on OCSSPs pursuant to art. 17(4)(a) should be modulated accordingly, e.g., requiring a lower remuneration.

One key criterion to assess compliance with the obligation in art. 17(4)(a) is that licensing terms “should be fair and keep a reasonable balance between both parties” and be subject to the payment of “appropriate remuneration” (p. 10; see also Recital 61). The Guidance takes this to mean, first, that “fair terms” are to be assessed on a case-by-case basis. In addition, a fair licensing deal means that OCSSP must be transparent about criteria used to identify and remunerate use of works, including as regards the use of content recognition technology. This addresses an important ask by rightsholders, who often complain about lack of detailed information in this context.

In conclusion, the Guidance states that for OCSSPs to comply with their best efforts obligation under art. 17(4)(a) they must accept to conclude licenses that are “offered on fair terms” and maintain “a reasonable balance between the parties”. Whether art. 17(4)(a) ends up in practice as a “best efforts” obligation or a more far-ranging obligation to license will depend on how these “fair terms” and “reasonable balance” are ultimately defined, and applies to different OCSSPs and types of content.

The situation is a bit clearer and perhaps less problematic for licenses that are negotiated with CMOs, which must comply with the requirements of art. 16 of the Collective Rights Management Directive. For that reason alone, it might be worth seriously exploring collective authorisation models for the right in art. 17.


To make sure you do not miss out on regular updates from the Kluwer Copyright Blog, please subscribe here.

Kluwer IP Law

The 2022 Future Ready Lawyer survey showed that 79% of lawyers think that the importance of legal technology will increase for next year. With Kluwer IP Law you can navigate the increasingly global practice of IP law with specialized, local and cross-border information and tools from every preferred location. Are you, as an IP professional, ready for the future?

Learn how Kluwer IP Law can support you.

Kluwer IP Law
This page as PDF

Leave a Reply

Your email address will not be published. Required fields are marked *