1 Taking Stock of the Past Ten Years

With a significant acceleration after a calm decade (2001–2011), which was mostly devoted to preparatory works and consultations, in the past ten years the European Commission (EC) tabled a set of important interventions, the last one being the already historical Directive on Copyright in the Digital Single Market (2019/790/EU, CDSMD). The CDSMD represented a shift in the approach to EU copyright harmonization, which, from the InfoSoc Directive (2001/29/EC) on, was characterized only by narrowly-scoped acts limited to fields that created significant problems to the correct functioning on the internal market and the pursuance of other Treaty goals.

The Orphan Works Directive (2012/28/EU, OWD) removed a long-standing obstacle to the preservation of cultural heritage by introducing for the first time a comprehensive mandatory exception to allow the digitization by cultural heritage institutions of printed, cinematographic and audio-visual works, phonograms and works embedded therein. To adequately protect rightholders, the Directive subordinated the declaration of orphan status to the performance of a diligent search and the registration of the work on the database OrphaNet, managed by the European Union Intellectual Property Office (EUIPO). While largely praised for the mandatory nature of the exception, which could overcome territoriality and create greater legal certainty and cooperation, the mechanism envisioned by the Directive soon proved to be too burdensome and complex to be handled by non-profit institutions. By 2018, only a bit more than 6000 works had been entered into the registry, thus evidencing how the new system was not sufficient to enable the mass digitization efforts the EU wanted to facilitate.Footnote 1

Two years later, the CMO Directive (2014/26/EU) intervened to improve the functioning of the internal market of protected works by harmonizing the governance, financial management and transparency rules of collective management organizations (CMOs), regulating the new independent managing entities (IMEs), and introducing multi-territorial licenses for the online cross-border distribution of musical works. The effects of the Directive have been assessed by two independent studies,Footnote 2 which highlighted the overall positive impact of the reform on the opening of the market to competitors,Footnote 3 on rightholders’ freedom of choice of and withdrawal from CMOs,Footnote 4 on their participation to the CMOs’ decision-making process,Footnote 5 on revenue collection,Footnote 6 and on the market for multi-territorial licenses (MTL), which has greatly benefitted from the creation of licensing hubs for multi-repertoire MTLs.Footnote 7 Although small rightholders might have suffered from this increasing centralization, niche national repertoires appear to be well represented, albeit not too competitively due to their relatively high prices.Footnote 8 Moderate concerns have been expressed about the application of different rules to licensing entities competing on the same market, and about the fact that the management of certain rights is still reserved to CMOs, which may penalize new entries (particularly IMEs) and the overall competitiveness of the system.Footnote 9 Similarly, the long-standing presence of large national CMOs still constitute a barrier for competitors – a situation made worse by the tendency for rightholders not to choose foreign organizations for the management of their rights.Footnote 10 This may weaken the positive pro-competition effects of the Directive. At the same time, however, it may also help keeping repertoires less fragmented, thus increasing the negotiating power of CMOs vis-à-vis large users, to the benefit of a wide range of rightholders.Footnote 11

Similar steps forward have featured the field of copyright and disability. 2017 was the year of the Marrakesh Directive (2017/1564/EU) and Regulation (2017/1563/EU), which introduced a mandatory exception for the benefit of visually impaired individuals, on the basis of the international obligations undertaken by the EU as a signatory of the WIPO Marrakesh Treaty. This cultural policy piece was flanked by a long-awaited consumer protection intervention in favor of subscribers of online content services with the Portability Regulation (2017/1128/EU), which obliged providers to enable users temporarily present in another Member State to access and use the same content they would be able to access and use in their country of residence, on the same range and number of devices, and with the same functionalities. The effects of the Portability Regulation, which were evaluated in 2022,Footnote 12 have been largely positive, with a remarkable uptake by consumers, their general satisfaction, and very few or no complaints by consumers organizations and authorities.Footnote 13 Only small and medium-sized enterprises (SMEs) operating in sectors that have not offered portability services before (audiovisual and sport) lamented some difficulties in complying with the new requirements, yet at no additional costs.Footnote 14 Also rightolders’ organizations welcomed the Regulation as an effective means to meet consumers’ expectations, noting that its impact on licensing practices and rightholders’ remuneration has been very marginal.Footnote 15 As to the Marrakesh Directive, the mandatory nature of its provisions has led to a great degree of convergence of national solutions, with limited discrepancies that do not affect the breadth of their scope.Footnote 16 According to the responses to the recent “call for evidence” launched by the EC on the effects of the Directive, most beneficiaries praise the greater number of works now made available to disabled individuals and the cost saving derived from the sharing of accessible collections between authorized entities.Footnote 17 While this represents a major achievement in the field, a long road is still ahead with regard to other forms of disabilities and particular types of works, for which exceptions under Art. 5(3)(b) InfoSoc are still heavily fragmented across the Union.Footnote 18

It then took several rounds of negotiations to finally produce the CDSM Directive, by far the most comprehensive act of harmonization after 2001, tackling some of the most pressing issues that have been discussed across two decades. The CDSMD marks a definite turn towards mandatory exceptions, overcoming the long-standing problem of their territoriality and fragmentation, and intervening in key areas such as TDM, preservation of cultural heritage and digital education. It provides licensing and exceptions schemes to increase the availability of out-of-commerce works. It introduces the tool of extended collective licensing into EU law, dispelling any doubt as to their compatibility with EU law indirectly cast by the Court of Justice of the European Union (CJEU) in Soulier and Doke.Footnote 19 It takes position on very controversial issues such as the grant of a new related right to press publishers and the attribution of direct liability to online content-service providers for infringing content posted by their users (with related preventive content-filtering obligations). Last but not least, it strongly brings authors back to the main stage after years of industry-oriented legislation, introducing copyright contract law provisions that aim at tackling the unbalanced bargaining power between authors and publishers.

It is still too early to evaluate the impact that the CDSMD will have on the digital single market of protected works and the balance between the conflicting interests at stake. However, commentators have already abundantly highlighted its strengths and flaws, advanced specific reform proposals and, more generally, underlined several challenges and problems affecting EU copyright law that have still remained unaddressed.Footnote 20 This revived debate and conspicuous wave of policy recommendations, also triggered by the external intervention on copyright-related matters by the EU efforts on platform regulation and the data economy, stand in stark contrast with the fact that copyright has disappeared from the Union’s legislators legislative agenda and debate, and this will likely be the case for some years to come – as happened after the InfoSoc Directive.

But is it really true that there is not now and will not be a policy agenda for EU copyright?

2 A Copyright Policy Agenda for the Next Six Years

Looking at the most recent recommendations that have been advanced by renowned groups of European academics,Footnote 21 which are fully to be agreed with, this does not seem to be case.

Aside from long-term goals, which may not appear pressing under a realpolitik agenda planning, key matters that require prompt attention in the near future in order to optimize the current harmonization status of EU copyright, achieve a better balance between conflicting interests, and make EU copyright law fully fit for the digital and data economy can be summarized in three categories: (1) consolidation of the acquis, to solve inconsistencies and lack of conceptual clarity, tackle the weak reception of CJEU’s landmark doctrines, and streamline the overlap of conflicting regimes; (2) new areas requiring harmonization to pursue Treaty goals; and (3) the need for monitoring, impact assessment and feasibility studies in specific sectors.

Consolidation of the acquis. The first group of recommendations point to the need to consolidate the current acquis with interventions that aim at removing inconsistencies between rules and definitions set by subsequent acts, specifying unclear and debated concepts, ensuring the national operation of key CJEU doctrines, which are often disregarded by national courts to the detriment of legal certainty and solving conflicts created by the presence of overlapping regimes. While examples are many, some issues have proven to be more pressing than others in light of the problems they created in EU and national case law.

Inconsistencies that are particularly challenging are the different definitions (and thus) scope provided for similar notions, such as the right of reproduction in copyright versus the right of reproduction in related rights.Footnote 22 Similarly, due to the “sedimentary” nature of EU harmonization, we have different rules in different directives regarding similar exceptions. In this context, the most glaring examples are the disability exception (Marrakesh vs InfoSoc); the digital vs general education exception (CDSMD and InfoSoc); and, most recently, the parody, quotation and pastiche exceptions in the InfoSoc Directive and in Art. 17(7) CDSMD, which are subject to a different regime (overridable vs mandatory).Footnote 23

Lack of conceptual clarity that keeps on triggering legal uncertainty and calls for an urgent solution involves basic pillars of EU copyright law as set in the InfoSoc and Database Directive (96/9/EC), such as the boundaries of protected works, the notions of exhaustion and lawful user, the scope of the right of communication to the public, the breadth of the notion of temporary reproduction, the requirements of exceptions such as quotation and parody, the notion of “substantiality” of the investment or of the part of a database extracted and re-used.Footnote 24 Similarly, the reversion right introduced by the CDSMD represents a unique possibility for new income for creators, new exploitation possibilities for investors and new access to the public, but the notion of lack of use, particularly in the digital environment, remains unclear as to its actual meaning. A clarification of the point, also through the publication of soft-law guidance, could help the functioning of the provision and its consistent application across the Union.Footnote 25

Another important aspect that has been highlighted by scholars is the lack of consistency or precision in provisions designating a single governing law to apply copyright rules, which triggers legal uncertainty in a highly territorial-based framework like the one characterizing national copyright laws. A consolidation of the current acquis may be the opportunity to revise connecting factors, seek consistency and clarity between specific provisions, and clarify the situation where the connecting factor may involve more than one jurisdiction (e.g. place of establishment or habitual residence).Footnote 26

A comprehensive mapping of the state of the art of copyright flexibilities across the EU has also highlighted the weak reception, by several national courts, of landmark CJEU doctrines that have reached a substantial degree of development (e.g. the fair balance doctrine or the principles enshrined in Deckmyn and in the 2019 trio).Footnote 27 Ensuring their uniform implementation is strongly needed to increase legal certainty across the Union and ensure that EU copyright law effectively reaches its objectives.Footnote 28 The same applies to the interpretation of Art. 5(1) InfoSoc on temporary reproduction, as well as Art. 5(3)(a) and (d) on illustration for teaching or scientific research and quotation.Footnote 29

In addition, some areas have been plagued since their onset by the uncertainty and overprotection created by the overlap of different provisions and regimes, which calls for a clarification by the EU legislator. The main examples here are the potentially concurrent applicability of the (EU) sui generis and (national) unfair competition law provisions in the field of databases, and the different anti-circumvention provisions applicable on software programs and on protected works in general.Footnote 30

New areas requiring harmonization to pursue Treaty goals. The realization of a well-functioning digital single market of protected works for the benefit of all stakeholders involved, the protection of fundamental rights and the fulfillment of EU cultural policy goals are still partially hindered by some flaws and gaps in the harmonization of EU copyright law. While several remarkable steps towards a more pervasive harmonization have been made in the past two decades, there are still matters which demand a relatively urgently resolution, due to the negative impact they have on legal certainty, cross-border exchanges and the four fundamental freedoms, and ultimately on the copyright balance. Some interventions could be limited to simple revisions of existing norms; others may require to legislate ex novo.

In the first category (revisions), the most pressing issue concerns exceptions. The shift in the approach towards mandatory limitations and exceptions (L&Es), marked by the CDSM Directive, has been only forward-looking. Existing exceptions, particularly those under Art. 5 InfoSoc, still remain optional and vague in language, as opposed to highly harmonized exclusive rights – and this despite the CJEU having repeatedly been invited to consider as indirectly mandatory L&Es that protect fundamental rights. A decisive intervention on the matter is strongly needed in order to overcome the negative effects of territoriality, with the country of origin principle being the smoothest and fastest way to reach an effective result.Footnote 31 This would be particularly important for exceptions linked to freedom of expression and artistic freedom (parody, caricature, pastiche, quotation),Footnote 32 especially in online settings, where the great divergences among national solutions create substantive obstacles for outlets operating cross-border.Footnote 33 Similar concerns suggest the need to update the informatory purposes exceptions to the new online information industry, including effectively new digital actors and online uses through a technologically neutral and purpose-oriented language.Footnote 34 In addition, the opportunities offered by AI and data infrastructures for research and innovation within the EU could be fostered by intervening on the research exception enshrined in Art. 5(3)(a) InfoSoc, which could be amended with further specifications – distinguishing, for instance, between teaching and research purposes and clarifying the application requirements – and made mandatory in order to become AI- and data-proof and be consistently applied across the Union.Footnote 35

In the second category (new legislations), one of the most pressing needs is that of aligning EU copyright law with EU policies on Open Access and Open Science, as also recently indicated by the Commission. Necessary steps may be the introduction of a mandatory research exception modelled on Art. 5 CDSMD, to solve the fragmentation affecting national implementations of Art. 5(3)(a) InfoSoc, and of an EU-wide secondary publication right limited to OA via self-archiving, not overridable by contracts and attributed to authors.Footnote 36

The advent of the Marrakesh Treaty and its implementation has shed light on the fragmented patchwork of national disability exceptions and on the challenge copyright still poses for disabled people beyond the visually impaired. Complying with the obligations set by Art. 30 of the UN Convention for the Rights of People with Disabilities requires Member States to remove all obstacles posed by intellectual property rights (IPRs) to the fulfillment of the right to culture of disabled individuals. This calls the EU to put on the agenda a broader intervention on disability L&Es, making mandatory, and regulating in greater detail, the InfoSoc exception following the example of the Marrakesh Directive.Footnote 37

Another area where Member States show divergences is the treatment of conducts that formally fall under an exclusive right, but do not conflict or compete with the normal exploitation of the work. Such “transformative uses” are shielded by some national courts and banned by others. The European DSM and its cultural and creative industries may benefit from a clarification on the matter, based on an impact assessment of the potential economic and non-economic effects of a new transformative use exception.

Still in the field of L&Es, the CDSMD brought the welcome addition of two text and data mining (TDM) exceptions, necessary for the achievement of a number of EU Treaty goals (fair access to data and technology, pluralism, inclusiveness, competitive AI business models). However, the implementation of Arts. 3 and 4 CDSMD will need to be strictly monitored to assess its effects and verify whether some of its narrower aspects (e.g. research purposes by limited beneficiaries under Art. 3, the possibility to reserve the right to TDM in Art. 4, and the filter of lawful access under both provisions) are really at risk of frustrating the objectives of the two exceptions.

Along with a consolidation of the acquis on the issue, a clearer delineation of the boundaries of the public domain with clear-cut standardized exclusionary rules is now of utmost importance vis-à-vis the advent of AI, the surge of the data economy and the development of new forms of creativity. Article 35 of the Data Act, with the exclusion of IoT machine-generated data from the database sui generis protection is a first step in this direction, but many more are still needed to ensure a proper harmonization and avoid overprotection and distortions in the copyright system.

As to the protection of performances, in light of the impact of AI music outputs, it is becoming increasingly important to understand whether and how national legislations regulate the matter; that is, whether or not they subordinate the granting of related rights to the performance of “works”. Since this qualification is currently in doubt for AI-generated creations, it is advisable to consider an EU intervention on the requirements for the attribution of related rights in the case of performance of AI-based works. At the same time, the absence of clear and univocal economic evidence that supports their introduction, no new rights (existing or sui generis) should be granted on computer-generated works, in order to avoid negative impacts on incentives to human creativity (an approach also labelled as “wait and see”).Footnote 38

Areas requiring monitoring, impact assessment and feasibility studies. Lastly, it would be advisable to deepen the understanding of a number of regulated and non-regulated phenomena within the DSM to better inform future decision-making processes.

First, it is important to acquire more data on contractual practices within the creative industries in order to effectively monitor the impact of the new provisions on copyright contract law introduced by Chapter 3 of the CDSMD. The current lack of real knowledge on the matter makes it almost impossible to verify whether the recent reform is of real help in rebalancing the contractual power of parties involved, or if future interventions are still needed to ensure proper authors’ remuneration.Footnote 39

It may also be worth exploring the way in which creative industries are regulating rights attribution. A proper understanding of artistic and business practices in the field of AI-generated creations is of key importance to guide the development of the market, and to define the most appropriate regulatory tools to ensure balanced outcomes and the uptake of each intervention by all concerned parties.Footnote 40

Another area where evidence is mostly lacking is that of TDM practices and its impact on the EU data economy and regulatory competition between legal systems. Particularly after the implementation of the CDSMD, it is of utmost urgency to understand whether the exceptions envisioned by the EU legislator under Arts. 3 and 4 CDSM have caused any change in the behaviors of market actors vis-à-vis the development of AI applications within or outside the Union. The restrictions imposed by the two new exceptions may incentivize firms to train models in legal systems that are more flexible towards TDM practices, or to import pre-trained models from outside the EU to avoid incurring in high licensing costs. Should that be the case, this may slow down the development of the EU AI industry – an effect that the Union may want to avoid. An impact assessment of the effect of Arts. 3 and 4 CDSM on these practices is urgently needed to inform future policy actions in the area.Footnote 41

With regard to the debate on the regulation of copyright matters in AI, it is also advisable to assess the operation of provisions that are already in force before deciding on new interventions. In this context, for instance, it has been correctly suggested to scrutinize the presumption of authorship and ownership under Art. 5 of the Intellectual Property Rights Enforcement Directive (IPRED), focusing on areas where a declared absence of authors could cause market players save some copyright-related costs (e.g. royalty payments to authors) and where the declared presence of authors could instead create revenues (e.g. based on copyright protection of AI outputs). Assessing how Art. 5 IPRED may help users of AI systems disclose and retain authorship over works generated by AI with their contribution may offer clear evidence on whether or not a legislative intervention in the field is effectively needed.Footnote 42

Another area that requires additional investigation is the streaming market. After the entry into force of the DMA, it is still debatable whether the remedies currently available are enough to allow creators and users to easily switch platforms by transferring their full network data, thus increasing the contractual power of creators and their opportunities for remuneration. The implementation of interoperability and transfer of copyright-related data should be subject to a feasibility study on potential legislative interventions, aimed at pursuing at best the CDSM and Data Package objectives.Footnote 43

Commentators have also highlighted the need for more research on the extent to which existing copyright rules apply to copyright content monetization by online content-sharing service providers, on the impact of copyright content moderation and recommendations on access to culture and cultural diversity,Footnote 44 and on the need for transparency and access to data held by platforms for researchers, also vis-à-vis trade secret protection, with a view to proposing legislation if need be.Footnote 45

While some actions may require careful scrutiny and long consultations, other interventions could be implemented swiftly and not trigger controversies among stakeholders and Member States. Ordering each matter in a specific list of priorities goes beyond the task of scholars and is necessarily subordinated to considerations that are proper of policymaking activities. Whatever the decision of the next Commission will be, it is beyond doubt that postponing or deleting EU copyright from the policy agenda for the upcoming five years would carry unintended negative consequences for the proper functioning of the digital single markets, curtail remuneration opportunities for stakeholders, and continue compromising the full enjoyment of users’ fundamental rights – a result that the new Commission may not want to be remembered for in the following decades.