While at the beginning of the week only three Member States had fully implemented the Directive (the Netherlands, Hungary and Germany), and could therefore be scored on all performance levels, it is already possible to track the level of activity across the board. As more Member States reach the finish line, we will attribute final scores and throw them into the contest. 

The first, second and third places (so far!)

So far, Germany is the front runner: the Federal Ministry of Justice and Consumer Protection held a transparent and inclusive discussion, which lasted for more than a year, and set a high standard for protecting user rights against overblocking. Hungary is in second place, in part due to the bonus point it got for fast-tracking the implementation of the new digital education exception, during the outbreak of COVID-19, having created room for remote teaching while educational institutions were closed. The Netherlands have been the first out of the door, with a draft text ready for an online consultation less than a month after the publication of the Directive, but the Dutch government failed to demonstrate its commitment to protecting user rights in the implementation, pushing it to the third place so far (with the possibility to still earn some extra points, if the Minister of Justice decides to make use of the power that received in the implementation law, to provide further rules for the application of Article 17).

France and Denmark, which have rushed to implement on time only the provisions that strengthen the position of creators and right holders, have been scored for the implementation of Articles 15 and 17, but will only officially enter the contest once they have implemented the remaining parts of the Directive.

Skipping the parliamentary debate

At this point, all Member States (except Portugal) have, in some way or another, initiated the legislative procedure, but some processes have been far from transparent or inclusive. In France and Italy, the Parliament delegated the legislative powers in the government, meaning that those countries will skip a central stage of the democratic process, which is the parliamentary debate and vote over the concrete implementation proposal put forward by the government. In France, where the Ministry of Culture went through the implementation of Articles 15 and 17 without providing any opportunity for stakeholders to share their views and concerns about those provisions, no public consultation is expected for the remaining parts of the Directive. In Italy, the Ministry of Culture is said to be planning to, at least, run a public consultation once its draft decree is finalized.

In Ireland, the Department of Business, Enterprise and Innovation held a series of public pre-draft consultations, each focused on a different part of the Directed and all carefully prepared, but a few days ago announced that the Directive was going to be transposed into Irish law by way of regulations contained in secondary legislation, without submitting the actual draft law to public discussion and without further parliamentary debate.

In Spain, a first round of public pre-draft consultations was organized at the end of 2019, but since then the Ministry of Culture went silent, leaving meeting requests by civil society representatives unanswered, while at the same time meeting with collecting societies and other industry representatives. Most recently, there have been rumours of a Royal Decree that would give the Spanish government the power to also implement the Directive without any further parliamentary deliberation.

Portugal is the slowest country so far (it has not published any draft text and has not even organized a public consultation), and civil society representatives have also been treated less favourably by the Ministry of Culture than representatives of rightholders: the latter have been invited to provide feedback on the implementation and have been granted the opportunity to meet with the Minister of Culture, while meeting requests sent by civil society organisations (including by COMMUNIA) to the Ministry have been left unanswered.

Meaningful consultation procedures: some more, others less

Fortunately, the procedures in that handful of countries seem to be the exception, rather than the rule. The vast majority of Member States attempted to give all stakeholders, and not only a selected few, opportunities to participate in the discussions. Some countries decided to consult stakeholders at the very beginning of the process, before embarking in the drafting process, others kept discussions alive throughout the process, and others yet reserved most time for dialogue after releasing the draft.

Of the 19 Member States that organized pre-consultation processes, 8 held formal consultations, 8 opted for setting up working groups or entering into other forms of dialogue with stakeholders on the provisions of the Directive, and only 3 organized both a consultation and meetings with stakeholders before starting the drafting process. After the release of the draft, 7 countries organized formal consultations, 6 engaged in dialogues with stakeholders, and 2 (Germany and the Netherlands) opted by running formal consultations and engaging also in constructive dialogues with the stakeholders.

So far, Greece, Italy, Finland, Malta, Norway, Portugal, Poland, Spain, Sweden are the only countries that have not publicly released any draft implementation text. France and Denmark have yet to release proposals for the parts of the Directive that have not yet been implemented.

It is safe to say that the Eurovision DSM contest is far from over, and that the results can still change dramatically. We will keep tracking the process and updating not only each country’s scores, but also each country’s page in our implementation tracker, where detailed information and documentation about each process can be found.

The post Eurovision DSM Contest: the once in a decade copyright reform contest appeared first on COMMUNIA Association.

While the Dutch government has reacted to criticism from civil society and members of Parliament by fixing some of the most obvious shortcomings of its implementation law, the 2nd chamber of the French Parliament has adopted a law that gives the French government the power to implement Article 17 (and the rest of the provisions of the DSM directive) however it sees fit. 

Netherlands: a course correction

Back in July of 2019, the Netherlands were the first country to propose an implementation law of the DSM directive. Somewhat surprisingly (the Netherlands had been one of the most vocal opponents of Article 17 in the Council) the proposed implementation law did not make any efforts to protect user rights and omitted most of the user rights safeguards contained in the final version of Article 17. After the proposal was sent to Parliament in June this year, together with Bits of Freedom and others we pointed out these shortcomings to the members of the legal affairs committee. Yesterday, in response to questions from members of the legal affairs committee, the government conceded that its original implementation proposal was incomplete and added the missing user rights safeguards to the proposal for an implementation law.

The updated implementation law proposal now includes all elements of Article 17 of the DSM directive. While it is disappointing to see that the Dutch government is sticking closely to the text of the directive (and not following Germany in exploring the flexibility offered by the directive), there is still room for improvement. The implementation proposal contains a clause that would allow the government to issue more detailed rules on the application of the article by means of an administrative order. In its response to the Parliament, the Dutch government has made it clear that it intends to use this clause to implement the Commission’s upcoming guidance on the implementation of Article 17, as long as such guidance will not further limit user’s rights. In this context the Dutch government has also made it clear that it supports core elements of the Commission’s proposed guidance. The Minister of Justice writes:

I follow the European Commission, which, on the basis of the stakeholder consultations carried out so far, is of the opinion that the application of automatic filters should be limited to those cases where there is a likelihood of infringing uploads (‘likely infringing’).

France: soldiering on

As we have reported earlier this week, the French government has made it clear that it is not interested in implementing any additional user rights safeguards. This is reflected in its approach to implementing Article 17 into national law. After having presented an initial proposal for an implementation law in December 2019 – one that also omitted most of the user rights safeguards contained in the final version of Article 17 – the French government changed its approach. In the face of possible delays because of the COVID-19 emergency, the French government added provisions to the law “DDADUE” that would authorise it to implement the provisions of the new copyright directive by decree (and without further parliamentary discussion). This so-called “DDADUE” law was approved by the Senate in July of this year and yesterday the Assemblee Nationale approved it as well. Once signed into law by the President, the government will have six months to implement the provisions of Article 17 into the French intellectual property law. 

In theory, the French government could use this time to wait for the Commission’s guidance and to add the missing user rights safeguards, but given the French government’s strong alignment with the position of rightholders and commitment to a rapid implementation, it seems more likely that it will go ahead and implement the law as soon as possible, without any substantial changes from the version presented in December. Given that the resulting law will likely lack important user rights safeguards contained in Article 17, the legality of such a rushed implementation of Article 17 remains highly questionable. 

And while France will likely be the Member State which can make the (dubious) claim to fame of being the first Member States to implement Article 17, the rest of the Member States would be much better advised not to look at France for guidance. The one thing that all of them should learn from these concurrent implementation efforts is that Parliamentary scrutiny contributes to better legislative outcomes. While the Dutch Parliament asserted its right in order to fix obvious flaws in the government’s proposal, the French Parliament has passed on this opportunity, severely undermining the legitimacy of the French implementation.

The post Implementation update: French Parliament gives carte blanche, while the Netherlands correct course. appeared first on COMMUNIA Association.

These first implementation proposals are coming from a main proponent of Article 17 (France) and one of the most vocal opponents (Netherlands), and allow us to get a first impression of how Member States across the EU are likely going to deal with this controversial article. Irrespective of the different positions by France and the Netherlands during the directive negotiations, the implementation proposed by both Member States do not diverge much from each other.

Both the laws stay very close to the text of the directive: The French implementation largely follows the order of the different sections of the directive via two nearly identical articles, one dealing with copyright (L137) and the other dealing with related rights (L219). The Dutch implementation law follows its own structure and introduces 3 articles (29c, 29d and 29e) that deal with copyright and one article (19b) that declares these articles to also apply to neighbouring rights.

Selective implementations

The general approach chosen by both legislators is to transpose the text of Article 17 (paragraphs 1-6 and 8 and the definition of online content-sharing service providers (OCSSP) from article 2(6)) as closely as possible into their national law.

Neither of the legislators transposes paragraph 17(7) (that introduces crucial safeguards protecting uses under exceptions and limitations) and both only transpose parts of paragraph 17(9) (that imposes an obligation on OCSSPs to operate a complaint and redress mechanisms for users in the event of disputes over the takedown and staydown procedures). None of the legislators provide any further guidance on how platforms are supposed to meet the requirement to make “best efforts to obtain authorisation” from rightholders.

There are also a small number of notable deviations from the text of the Article in the Directive:

• The Dutch law includes a clause that allows the legislator to impose additional rules for implementing the article, via an administrative order. This provision seems designed to take into account the outcome of the EU-level stakeholder dialogue and means that it is difficult to fully assess the Dutch implementation at this stage.
• The Dutch law does not transpose the first sentence of paragraph 8 (the provision dealing with general monitoring) while the French law contains a clause that states that OCSSPs “shall act only on the basis of relevant, necessary information or notifications received from rightholders”.
• The French law explicitly includes “indirect commercial advantage” in the definition of OCSSPs (in the directive “indirect” is only mentioned in a recital).
• The French law provides that the Conseil d’Etat can issue a decree on the modalities for defining “large amount of works” in the definition of OCSSPs.
• The French law includes language that implies that OCSSPs also need to obtain authorisation for acts of exploitation other than acts of communication/making available to the public (this would result in a substantial expansion of the scope of the Article).

User rights

As noted above both proposed implementation laws only implement parts of the user rights safeguards contained in paragraphs 17(7) and 17(9) of the directive.

With regards to paragraph 17(7), both proposed laws ignore the first part with states that the collaboration of OCSSPs and rightholders, in line with paragraph 17(4), shall not lead to the prevention of the availability of content that does not infringe copyright. As a result, neither of these laws introduces any requirements on OCSSPs and/or platforms to ensure that non-infringing works are not blocked by overly aggressive upload filters.

None of the implementation laws contain any provisions based on the second part of Art 17(7), which requires that users “are able to rely on” quotation and parody exceptions. While both France and the Netherlands have existing quotation and parody exceptions in their copyright acts, it seems questionable if this also means that platform users can effectively rely on them in all circumstances, since both countries copyright laws allow technological and contractual override of these exceptions.

Both legislators selectively implement elements of Article 17(9). Most notably neither of them transposes the first part of the penultimate paragraph (“This Directive shall in no way affect legitimate uses, such as uses under exceptions or limitations provided for in Union law”) and stick to mechanically implementing the elements of the complaint and redress mechanism contained in the first two paragraphs of 17(9). However, both of them limit the requirement for human review of decisions to disable access to or remove uploaded content to decisions that are challenged by users.

There are also a few relatively small differences between the two implementation laws:

• The Dutch implementation law does not transpose the requirement that rightholders need to “duly justify” their requests (they only have to be heard by the OCSSP). The French implementation law only partially transposes this requirement, since it does not cover initial requests but only subsequent requests to block/remove after a user complaint.
• The French implementation law designates ARCOM (HADOPI successor) as the entity providing the out of court redress mechanism required by paragraph 17(9). The Dutch implementation law requires OCSSPs to make sure that users and rightholders have access to an impartial out of court redress mechanism without specifying an entity.

It is worth noting that this selective approach to implementing the user rights provisions contained in Article 17 has already been rejected by the European Commission. In its answer to a parliamentary question by MEP Marcel Kolaja, the Commission has made it clear that Member States need to impose on OCCSPs the user rights safeguards contained in the directive  as part of their national implementations. The current French and Dutch proposals fail to do this.

Assessing the proposed implementations against COMMUNIA objectives

In October 2019 we formulated five objectives for minimising the harm caused by Article 17. These are:

1. A targeted definition of OCSSPs.
2. Ensuring wide availability of licenses.
3. Removal requests via a centralised public database.
4. Broad spectrum of exceptions and limitations.
5. Strong default protection of user rights.

Both proposed implementations fail to meet most of these objectives:

1. Neither the French or the Dutch definitions of OCSSP further narrow down the type of services that will need to comply with Article 17 of the Directive. Notably they do not make use of the additional criteria contained in recital 62, which clarifies that the aim of the Directive is to target ad-supported engagement-driven service providers that play an important role in the online content market by competing with other services for the same audiences.
2. Neither of the proposed implementation laws include any provisions on the modalities of licensing OCSSPs for the use of works uploaded by their users. In the absence of such provisions, direct licensing and voluntary collective licensing are the only possible options.
3. Neither of the proposed implementation laws include any provisions regarding the modalities that rightholders have to follow for communicating the relevant and necessary information about works that they want to be blocked/filtered to OCSSPs.
4. Both France and the Netherlands already have quotation and parody exceptions in their national acts as required by paragraph 17(7). While the Dutch copyright act also has an incidental inclusion exception, the French act does not have such an exception. In addition both laws likely do not meet the requirements that platform users can effectively rely on these exception in all circumstances as both copyright acts allow technological and contractual override of these exceptions
5. Neither of the proposed implementation laws impose any requirements on OCSSPs to ensure that user rights (uses under exceptions) are not impacted by any preventive measures implemented in line with paragraph 17(4). Neither of the proposed implementation laws afford default protection to uses that are potentially covered by an exception or limitation.

In the case of the Netherlands the above analysis is not final since the ability to provide additional rules via administrative order means that some of the missing aspects could be introduced by the legislator after the implementation law has been adopted.

Conclusion

The Dutch and French implementation law proposals show that national legislators cannot be trusted to implement Article 17 in a balanced way that includes the hard-fought user rights safeguards that are part of the final compromise that was found between the EU legislators after more than two years of bitter discussions.

Under pressure from rightholders, national legislators are focusing their attention on those parts of the Article that require rightholders and platforms to cooperate with each other to either license or filter. Without the user rights safeguards in the latter part of Article 17, such cooperation will undermine users freedom of creative expression. Not implementing the user rights safeguards is a glaring act of disregard of the outcome of the EU legislative process, and will need to be challenged once these legislative proposal reach the national parliaments.

The post Article 17: Both French and Dutch implementation proposals lack key user rights safeguards appeared first on COMMUNIA Association.

We would like to provide here an overview of the Dutch proposal to implement locally the new EU educational exception (article 5 in the final version of the Directive). This is the beginning of our effort to track how countries across Europe will implement, over the coming two years, this mandatory exception to copyright for educational purposes.

We are in particular interested in three issues that have been our concern during the legislative debate on this exception:

• Are Member States introducing the controversial article 5(2), through which they have the option to make the exception no longer applicable and available to educational establishments if “suitable licenses are easily available on the market (what we call the issue of “license priority”);
• What is the scope of the exception:
  • How are educational institutions and staff defined?
  • Will the educational community be able to rely on email, cloud services and other password-protected environments, or will these not be considered “secured electronic environments” under the exception?
  • Will Member States define a priori the extent to which a work can be used, leading to different quantity limits in different countries, or will they let practice and courts (relying on the three-step test) define what is balanced in a given situation?

Choices made on these issues will determine, how broadly – or narrowly – can the exception be depended on. Taken together, they will also create either a harmonized or fragmented legal landscape for teachers and learners across the Union.

• Are Member States changing remuneration rules for educational uses? Currently, 17 Member States do not remunerate most or all of the used permitted under their existing education exceptions – will this change with the new exception?

The Dutch proposal is a simple amendment that adds two paragraphs to the existing educational exception (article 16 of the Dutch Copyright Law). In relation to our issues of particular concern, we note that the Dutch government:

• decided not to use the article 5(2) backdoor to hack the new educational exception and make it partially or fully not applicable in the Netherlands, which we applaud (because we believe – along with the CJUE – that users should have the right to benefit from the copyright exceptions that were created for their benefit at all times, and not only when there are no market options for them to get a license for those minimum uses that are protected by the exception);
• could do more to provide as broad a scope for the exception as possible, within the boundaries set by the Directive;
• has proposed not to change its approach to remuneration – use of content under the exception requires fair compensation (art. 16.1.5°).

Additionally, the proposal includes an explicit provision against contractual override (art. 16.6), which implements another important element of the new EU educational exception.

We will be working with our Dutch partners in the consultation phase, both to provide feedback on the government’s proposal, and to monitor other responses to the proposal. The consultations are open until 2 September 2019.

The post The Netherlands leads way with implementation of the new educational exception appeared first on COMMUNIA Association.

COMMUNIA’s 2012 analysis of the directive showed that it was bound to be a train wreck. A preliminary comparative study of the situation in the UK, the Netherlands, and Italy undertaken by the EnDOW project reveals that the national implementations of the directive across Europe do not provide the much needed solution for the problem of orphan works.

Under the directive, cultural heritage institutions are allowed to publish works online for viewing (not re-use) after a ‘due diligence search’ has been performed, recorded, and submitted to the orphan works database at OHIM. Works that have been registered in this database can then be digitized and made available online under an exception to copyright. So far the project only published its initial results, but we can already see that this piece of legislation will most likely not contribute to large-scale use of orphan works by Europe’s Libraries, Museums & Archives.

The main reason for this is that the diligent search requirements established by the directive have been implemented by member states in such a way that the cost of undertaking a diligent search is prohibitive. The study collected over 210 sources, databases, and registers that need be checked in diligent searches in the UK alone. Researchers from Italy found 357 possible databases and registers. Of the 87 identified sources in the Netherlands, 40 were not freely accessible, and 36 of these required personal contact or a physical visit to an institute. Since the legislation requires cultural heritage institutions to be diligent, they need to check each and every source to be covered by the limited exception provided by the directive.

These results illustrate that the EU approach to orphan works is unreasonably complex and won’t adequately address the problem it’s trying to fix. This is further shown in the actual number of orphan works available through the OHIM Orphan Works Database, which currently only shows 1,435 registered works. More than half of them are in the collection of the Dutch EYE Film Institute (which has worked on rights clearance for these works since at least 2008).

The preliminary results of EnDOW provide evidence that the European Union has failed in this attempt to provide much needed digital access to Europe’s cultural heritage. Given that the Orphan Works Directive does not help with mass digitisation projects, this means that there is a continued need to provide legal mechanisms that allow cultural heritage institutions to make works in their collection available online.

Note: This contribution has been written by Maarten Zeinstra. Maarten is technical advisor to EnDOW. The ideas expressed in this post should not be attributed to EnDOW.

The post Research: Orphan Works Directive does not work for mass digitisation appeared first on COMMUNIA Association.

The Anne Frank Stichting announced plans to publish an edition of Anne Frank’s texts online after the presumed expiration of the copyright on January 1, 2016. In response, the Anne Frank Fonds sued the Stichting over what it considered unauthorised reproductions of Anne Frank’s writings. The reproductions had been made by the Stichting as part of its preparatory research for the on-line publication after the new year. Initially, this seemed to be an attempt by the Fonds to thwart or delay the Stichting’s plans for an online edition.

However, during the course of the legal arguments it became clear that under Dutch law (which governs uses made by the Stichting), Anne Frank’s original writings would not enter the public domain in 2016. This is due to a transitional rule in the Dutch copyright act which states that works posthumously published before 1995 will retain copyright—in this case large parts of the original writings will only expire in 2037.

While this means that the Stichting had to shelve its plan to publish an online edition, the Fonds continued to press charges related to the reproductions (XML-TEI files) made by the Stichting in order to carry out its textual and historical research. The Stichting was sued alongside their research partner the Dutch Royal Academy of Science (KNAW). Both upheld the position that it did not require permission for making reproductions solely intended to enable its internal scholarship, claiming that copyright law should not be used to thwart scientific research.

On the 23rd of December the District Court of Amsterdam handed down its ruling in the case. After establishing that the writings of Anne Frank are indeed protected by copyright (and, in the Netherlands, will continue to be protected for the foreseeable future), the court also ruled on the legality of the research reproductions made by the Stichting.

While the court dismissed arguments that the creation and use of these reproductions were covered by a number of exceptions and limitations to copyright, the court did agree with the claim that the requirement to obtain permission from the rights holder for making such copies is in in conflict with the freedom of scientific research as established by article 13 of the Charter of the Fundamental Rights of the European Union. In its ruling (Dutch, translation mine) the district court argues (emphasis added):

From [the previous arguments] it follows that the creation of the XML-TEI file that has been made available to third parties constitutes an infringement of the copyright held by the [Anne Frank] Fonds. It needs to be judged if the circumstances of this particular act provide a reason to reject the demands made by the Fonds because this would, in the light of the principle of proportionality, put unreasonable restrictions on the freedom of scientific research. […]

It goes without saying that in order to carry out proper textual scientific research the researchers must have access to some copies of the texts that are being researched. Without these reproductions it is impossible to access the source materials which makes the research virtually impossible. This includes the XML-TEI file produced by the Huygens ING [Institute]. After all, this file has been created […] for the sole purpose to carry out scientific research.

The Fonds has only broadly stated that it does not have to tolerate everything that happens with the texts. Insofar as the Fonds tries to obtain control over what research should take place or not, this is not a right that is protected by copyright.

It is also clear that the infringement of the copyright of the Fonds taking place as part of the research does not extend beyond the provision of only a few reproductions of the works, and to a limited number of researchers directly involved in the research. The copyright infringement thus has minimal impact.

Under these circumstances, the court concludes that enforcement of the copyright by the Fonds is subordinate to the fundamental right of the Stichting et al on her freedom of scientific research.

Anyone who is familiar with the current discussion about the copyright status of text and data mining will quickly recognize that this case—which started as a dispute about the length of copyright protection—offers some valuable insights into the legal status of text and data mining in Europe.

The actions of the Stichting and the KNAW (creating a machine readable version of the text (the XML-TEI file)) are an excellent example of text mining. Research organisations and research libraries have long claimed that the making of reproductions of works that happens as part of the process of text and data mining should not require permission from the rights holders as long as the researchers have legal access to the works in question.

The court supports this line of reasoning by recognizing that requiring permission from the rights holders before machine readable reproductions can be made would make TDM-based research ‘virtually impossible’. In addition, by tying the issue of text and data mining to the freedom of scientific research, the court provides a strong normative justification for the rationale that TDM should not require the permission from rights holders.

Since text and data mining is one of the issues that will be dealt with during the upcoming modernization of the EU copyright rules, we hope that European lawmakers will pay close attention to the reasoning of the court in this case. As the Anne Frank Stichting and the KNAW rightly point out, copyright law should not be used to thwart scientific research. Such an outcome is unfortunately a very real danger, given the approach presented by the European Commission in December.

The post What the diary of Anne Frank can tell us about Text and Data mining appeared first on COMMUNIA Association.