Here at the Kluwer Copyright Blog we are thrilled to have had the opportunity to ask Julia Reda MEP a few questions on the controversial Proposal for a Directive on copyright in the Digital Single Market (DSM Directive).
But first, some background. The original proposal was submitted in November 2016 by the Commission. The ordinary legislative procedure places the European Parliament and the Council of the European Union on equal footing, meaning that a separate process for assessing the proposal takes place within each institution.
At the Council, there was an agreement on 25 May on an amended version of the Proposal, which sets forth the Council’s position and provides the basis for its negotiating mandate (hereinafter the Council Version). At the EU Parliament, five Committees are involved in the discussions: IMCO (Internal Market and Consumer Protection), CULT (Culture and Education), ITRE (Industry, Research and Energy), LIBE (Civil Liberties and Justice), and JURI (Legal Affairs). The first 4 Committees have drafted and voted on their Opinions. The JURI Committee will vote on a compromise version on 20 June 2018.
The resulting text then requires confirmation by the entire Parliament in a plenary vote, thus forming the Parliament’s position. The vote is scheduled for either early July or September, and will be followed by so-called trilogue negotiations, i.e. closed-door negotiations between Parliament and Council, in which they will try to hammer out a compromise between the respective versions. Finally, this text must be approved in a final vote by the EU Parliament in Plenary.
Although the proposed DSM Directive has many controversial topics, we have decided to focus this interview on perhaps the two most controversial: the press publishers’ right (Article 11, on which see here, here) and the so-called “value gap” or “transfer of value” proposal (Article 13, on which see here, here, here, here).
We are very grateful to Ms Reda for sparing time in her busy schedule to do the interview, and here’s what she had to say.
1. What is the current state of the JURI Committee draft report to be discussed on 20 June, especially as regards Articles 11 and 13?
The rapporteur Mr. Voss has been unable to find a broad compromise on either of those controversial articles. In a recent interview, he even concedes being unsure if he has a majority. This is very unusual, a rapporteur is supposed to ensure broad support for a compromise text that all parties can live with. Instead, Mr. Voss has kept a hard line on both articles, fully endorsing direct liability for a broad range of platforms for their users’ copyright infringements, without any possibility to mitigate that liability, and maintaining the completely discredited neighbouring right with a few cosmetic changes. He will put his versions of Articles 11 and 13 to the vote against alternative compromises presented by other political groups, under my lead. It seems like Mr. Voss is relying on a very narrow majority in favour of upload filters in Article 13, consisting of his EPP group, the eurosceptic ECR, the liberal ALDE and the far-right ENF group, who together hold a one-vote majority in the committee. If he wins, it will be largely due to the support of the two members of the far-right Front National. On Article 11, the majorities are much less clear. It is a risky strategy that goes against the principles of European cooperation, where consensus is sought wherever possible in order to protect diverse interests.
Our alternative compromise proposals present a real compromise between the competing interests at stake. Under Article 11, publishers would be able to license their collections directly to websites and other users, based on a presumption rule that allows them to act on behalf of the individual authors who have contributed to their publications. If the copyright in a news article is infringed, the publishers will have standing in court under that presumption rule, which was first presented by former rapporteur Therese Comodini Cachia, and enhanced by the Estonian Council presidency. In our alternative Article 13, active platforms do have to conclude licensing agreements with rightholders under fair and reasonable terms, but at the same time there is a ban on automated upload filters as an enforcement tool, due to their gross inaccuracy and negative impact on fundamental rights.
2. You have criticized the proposed press publishers’ right for being a “link tax”. Could you explain why?
News articles are already protected by copyright whenever they constitute an intellectual creation of their author. However, short snippets from those articles such as the headline or a single sentence generally do not reach the threshold of originality. For example, there is no copyright in the sentence “Angela Merkel meets Theresa May”. Therefore, the use of such short snippets of news articles, for example to illustrate a hyperlink, is not a copyright infringement and does not require permission. This would be changed by the neighbouring right for press publishers. Unlike copyright, a neighbouring right does not require originality, because it does not protect an author’s own intellectual creation, but an investment by a producer, in this case the press publisher. Supporters of the neighbouring right have confirmed that it is the proposal’s very purpose to create an exclusive right over those small snippets like “Angela Merkel meets Theresa May” that are not yet protected by the journalist’s copyright.
The consequence of such a far-reaching new right would be that it would become impossible to reference a news article by its title or a short snippet when linking to it. This constitutes a significant barrier to freedom of information online and directly interferes with the common practice of referring to news articles by their title. The quotation exception will not be sufficient to ensure that such referencing remains legal, on the one hand because this exception has been implemented restrictively in some countries such as Germany, on the other hand because Mr. Voss’ text makes it clear that while Member States may apply existing copyright exceptions to the neighbouring right, they do not have to!
Why am I speaking about a link tax? The experience with the neighbouring right in Germany, where it was introduced in 2013, has shown that the publishers who have been lobbying for this new right do not in fact want to stop search engines or other websites from using such non-original snippets. In fact, they employ countless search engine optimisers and social media experts to make these snippets look as attractive as possible, to encourage more clicks. Their goal is to force large platforms and search engines to use their snippets and to pay for them. When Google shortened the snippets in Germany to comply with the new neighbouring right, the publishers first gave Google a free license to encourage it to keep using their snippets (of course they didn’t give a free license to the competition, thus strengthening Google’s market position), then complained to the German anti-trust authority that Google should be forced to use their snippets and pay for them. Unsurprisingly, the anti-trust body told them that no company can be forced to obtain licenses for content it doesn’t wish to use. Now publishers are trying the same strategy at a European level, with the aim of getting paid no matter whether companies actually value their snippets enough to use them in their hyperlinks, like a tax. Unlike a tax, unfortunately, the benefits of such a new right would not go to the public hands, but into the pockets of a few large media houses that try to use their influence over politics to change the basic principles of copyright and competition law.
3. In your view, what are the main problems and potential negative consequences of adopting Article 11, as worded in the Council Version or the current JURI draft report?
Aside from the general problems with the proposed neighbouring right I have described above, both the Council version and the JURI text create their own additional problems.
The Council has recognised that there must be some sort of lower threshold to protection, as the neighbouring right would otherwise even cover a single word and could lead to a privatisation of language itself – an impossible situation. Some Member States argued that the neighbouring right should only cover original snippets and hence be identical in scope to the copyright already enjoyed by journalists. While it is questionable what the added benefit of such an additional right would then be, at least it would largely avoid the negative impact on linking and referencing I previously described. However, other countries, led by Germany and put under immense pressure by German publisher Axel Springer, insisted that non-original snippets must be protected. The compromise found by the Bulgarian presidency combines the worst of both worlds: Member States are supposed to exclude “insubstantial” parts of news articles from the scope of protection, while each Member State may decide for themselves what constitutes an insubstantial part, either relying on originality, or on length, or both. This proposal of course is absolutely counter-productive to the stated goal of completing the digital single market. On the contrary it raises new barriers to cross-border trade and communications. Websites that want to operate EU-wide would probably have to apply the strictest interpretation of “insubstantial” to be able to operate legally in all countries. There would be no legal certainty for linking as it is entirely unclear whether a headline of a news article can be considered insubstantial.
The Parliament text, like the original Commission proposal, includes no lower threshold of protection at all. Only “acts of hyperlinking” are excluded from the scope, which may allow the use of a headline if it forms an integral part of the URL used to reference a particular news article, but whenever the URL does not include a snippet of the text, even the use of a headline to reference a news article would be forbidden. Another additional problem of the Parliament text is that it would not automatically apply all national copyright exceptions to the neighbouring right. Each national legislator would have to actively introduce a new exception to the neighbouring right from the list of optional exceptions to the InfoSoc directive, should they wish to apply for example the quotation exception. This would lead to immense additional complexity in the national copyright laws and create an additional barrier to cross-border application of copyright exceptions.
4. There has been rare academic consensus against Article 11. Furthermore, an open letter by more than 100 MEPs opposing the new right has been recently published. Do you think that such opposition could prevent the adoption of the right or force an alternative wording? If the latter, what version of this provision would you consider to be an acceptable compromise?
I think that the presumption rule championed by former Parliament rapporteur Therese Comodini Cachia and the former Estonian Council presidency is most likely to satisfy all sides of the debate. Publishers would be empowered to address actual copyright infringements of news articles and conclude new licensing agreements, albeit without the creation of a new right that would impinge on freedom of expression and information. That is why my political group Greens/EFA has put forward this presumption rule as an alternative compromise to the vote in JURI on June 20.
5. Moving on to Article 13, you and others have argued that this provision imposes mandatory upload and re-upload filtering that is inconsistent with the E-Commerce Directive and fundamental rights. Could you elaborate on this?
The text put forward by Mr. Voss applies to any platform that optimises user uploads for the purpose of showing those uploads to other users (with a very broad definition of optimising that even includes “displaying”), with very few specific carve-outs. That means almost any platform is covered, including for example Tinder, WordPress or TripAdvisor, even though their business models are in no way reliant on copyright infringement. Those platforms would be considered to be communicating to the public and would be removed from the hosting safe harbour of the e-commerce directive. That means they would be directly liable for all copyright infringements of their users. Unlike the Council text, there is no way for the platforms to mitigate their liability. In order to comply with the law, they would either have to preemptively get a license from every single rightsholder in the world for all the billions of copyright-protected works, which is impossible, because copyright arises automatically whenever a work is created and does not need to be registered, and because there are no collective management organisations representing all types of works, or they would have to prevent copyright infringements before they happen, which is also impossible, unless all works are censored and checked for copyright infringement before they become publicly accessible. Platforms will therefore have no other choice but to install automated filters to minimise the number of copyright infringements and reduce their liability. Filters will however not be enough to evade liability, as there are no filters that can accurately detect all copyright infringements. Mr. Voss tries to lull citizens into a false sense of security by stating that no general monitoring obligation shall exist wherever Article 15 of the e-commerce directive is applicable. Unfortunately, Article 15 of the e-commerce directive is only applicable to those providers that benefit from the hosting safe harbour under Articles 12 to 14 of the e-commerce directive, whereas any platform that falls under Article 13 of the directive on copyright in the digital single market, according to Mr. Voss’ own proposal, cannot benefit from the hosting safe harbour. In other words, his text confirms that the ban on general monitoring does not apply to platforms under Article 13, they are therefore required to install filters. It is very questionable whether this proposal is compliant with EU law, as the CJEU, in the Scarlett and Netlog cases, has grounded the ban on general monitoring in the Charter of Fundamental Rights.
Similar fundamental rights concerns have been raised in connection with the Council version of Article 13. In a letter to the European Commission, UN special rapporteur for freedom of expression and information David Kaye has recently rung the alarm bells, warning that “the restriction of user-generated content before its publication subjects users to restrictions on freedom of expression without prior judicial review of the legality, necessity and proportionality of such restrictions” and that the Council text “would also prevent a diversity of nonprofit and small content-sharing providers from potentially reaching a larger size, and result in strengthening the monopoly of the currently established providers, which could be an impediment to the right to science and culture”.
6. What, in your view, are the main problems and potential negative consequences of adopting Article 13, in either the Council Version or the current JURI draft report?
Article 13 in its currently proposed form would lead to the use of upload filters on the vast majority of websites regular Internet users use to interact with each other on a daily basis, and cause those websites that could not afford to install filters to shut down or withdraw from the European market as a consequence of excessive liability. The result would be a further concentration of the online platform market in the hands of a few multinationals that control upload filtering technology, primarily Google and Facebook. Under the Council text, these companies would even be clearly exempted from having to conduct license agreements, due to their use of filters. That means that authors would most likely receive even less money than they do today, where it is unclear to what extent Google or Facebook can benefit from the hosting safe harbour.
The impact on users’ fundamental rights would be dramatic: Copyright exceptions and limitations would become effectively meaningless in the online environment, because algorithms are incapable of distinguishing between legal quotations or parodies and copyright infringements. The filters would automatically remove any unlicensed use of copyrighted material and users would have to actively complain in the hope of being allowed to exercise copyright exceptions or limitations. Since exceptions and limitations do not constitute users’ rights and platforms would most likely filter material on the basis of their terms of service rather than the law, users would have a very low likelihood of success in their complaints. UN special rapporteur David Kaye has correctly pointed out this problem in his letter.
Another victim of Article 13 would be independent artists, as the proposal includes no safeguards against copyfraud, a widespread practice where primarily large companies such as TV stations wrongfully claim exclusive rights on all material included in their broadcasts, leading to the frequent takedown of original creations by smaller artists.
7. In your political efforts to oppose this Article 13, you have often labelled it as a “Censorship Machine”. What would you say to critics that argue such a moniker is excessive?
Even with today’s voluntary filters such as ContentID there is a plethora of well-documented cases of wrongful filtering of material. Because automated algorithms perpetuate the biases present in society, these filters tend to disproportionately affect vulnerable groups. With Article 13, we are putting in place a filtering infrastructure that will most likely be controlled by a very small number of companies that will develop the filters and perhaps even get access to the data streams from smaller platforms who cannot afford to run the filters themselves. It is naive to think that this censorship infrastructure will not be used for purposes other than fighting copyright infringement. The interior ministers of Germany and France have already called upon the European Commission to extend such filtering obligations to terrorist propaganda. It may be easier to convince politicians to support such dangerous censorship tools in the name of protecting authors, a sympathetic and vulnerable group. But once this system is in place, it will change the nature of online communication from a decentralised to a centralised system, vulnerable to government and business manipulation. Even if you trust the German or the French government with such tools, it is worth considering whether you can say the same about the governments of Hungary, Poland and Italy, or about Facebook and Google for that matter.
8. There has been significant opposition to Article 13 from academics, internet pioneers, civil society and other stakeholders, to name a few. On the other hand, you have yourself admitted that it may be justifiable to at least partially regulate certain platforms covered by Article 13. Assuming the criticism may allow the adoption of an alternative wording, what would be an acceptable compromise in your view?
First of all it is important to recognise that some online platforms have been treating authors unfairly, but that the use of filters has largely exacerbated the problem. If Google didn’t have ContentID, it wouldn’t be so easy to evade collective negotiations where authors can be represented by a collecting society. At the same time, it is also important to recognise that the vast majority of online platforms, while allowing users to upload copyright-protected material, do not exploit authors by doing so – just think of Wikipedia, ebay, Tinder, or the broad range of discussion forums that help diverse communities to find each other and organise around joint interests. A sensible response to Article 13 needs to firmly reject upload filters and instead focus on fair remuneration of authors. And it needs to clearly distinguish the first category of platforms like YouTube who monetise copyrighted content, from the second category, who cannot prevent uploads of infringing material, but in no way encourage it. A fair solution could take the form of a compensated exception, charging a levy to certain types of platforms and in return legalising the copyright-relevant acts of their individual users. Unfortunately, the copyright policy debate does not appear ready for such a radical solution. That’s why my political group Greens/EFA has introduced an alternative compromise to Article 13 that has already proven consensual in the LIBE and IMCO committees, which requires active platforms to conduct fair license agreements, but puts a clear ban on upload filters.
9. Finally, the debate on the proposal for a DSM Directive has been particular virulent, with seemingly little consideration of expert academic evidence in the final outcome of the Council or (some of the) Parliamentary Committees’ positions. What do you think are the reasons for this and how could the legislative process be improved in this respect?
There are several elements to this: On the one hand, reporting of EU politics needs to be intensified, to avoid the public debate taking place during national implementation, when the most important political decisions have already been made. Secondly, politicians need to immunise themselves against the significant lobbying influence of media companies that can put pressure on politicians not based on their economic significance, but on their power to sway elections. And finally, the media companies need to put in place a strict firewall between their business interests and their reporting, in order to prevent regulatory capture of this kind.