WORKSHOP

SUMMARY

MEET THE AUTHOR SERIES • 30 NOVEMBER 2018

The copyright directive and concerns for free expression, big data innovation and access to information

Summary by Laura Drechsler, FWO, Brussels Privacy Hub, LSTS, VUB

SUMMARY

On 30 November 2018, the Brussels Privacy Hub hosted a lunchtime debate on “The copyright directive and concerns for free expression, big data innovation and access to information”. The debate, moderated by Jens-Henrik Jeppesen (Centre for Democracy & Technology), featured contributions by Raegan MacDonald (Mozilla) and Aleksandra Kuczerawy (KU Leuven) and focused on the proposal for a new copyright directive by the European Commission that is currently undergoing trilogue negotiations. Within this proposal the focus was on three provisions that could potentially have long-term effects on freedom of expression, big data and freedom of information, namely Article 3 (text and data mining exception), Article 11 (press publishers’ right) and Article 13 (upload filters).

After a short introduction into the legislative path of the copyright directive by Jeppesen, the debate was started by Kuczerawy elaborating on the rules in the copyright directive for text and data mining (Article 3). She first explained that text and data mining was an analysis technique that was capable of dealing with a big amount of data (big data) and therefore useful for all areas of science, but also market research and journalism. As such text and data mining was also conducive to innovation as it could stimulate development of new research or products. However, due to the current legal regime that restricts text and data mining based on copyright reasons within the European Union (EU), the EU was at a competitive disadvantage compared to China and the United States (US) where text and data mining is more widespread and clearly allowed.

The problem according to Kuczerawy was that while data alone was not covered under a copyright regime, collections of data were included in the protection provided by the InfoSoc Directive and by the Database Directive. Current exceptions to these directives were not suiting the specific situation of text and data mining, as they are only suited for temporary use (so not for the more long-term use of text and data mining) or they are not mandatory, meaning that Member States might not grant this exception in their national implementation. The European Commission tried in its copyright proposal to remedy this situation by providing a more suitable exception for text and data mining with Article 3. However, originally, while the exception was mandatory, it was only so for non-commercial research, scientific research and research content where there was lawful access. This narrow scope would leave out application of text and data mining for market research, journalistic purposes, but also such research developed with commercial partners. Additionally, the focus on material with lawful access bore the risk of raising the costs of research and right-holders could start charging considerably for access. These issues were also spotted by the EU co-legislators, the European Parliament. According to Kuczerawy, the Parliament was currently considering an expansion of Article 3, to also include cultural institutions and commercial research on accessible works. Such an expansion is for the moment provided for in a mandatory exception, ensuring harmonisation across the EU.

As a second speaker, MacDonald, began her presentation by reiterating the legislative history of the copyright directive from the viewpoint of Mozilla. She explained, that initially Mozilla was very positive about a reform of the EU copyright regime as the one in place was outdated and created many issues, especially due to the lack of harmonised exceptions, as it dated from before the ubiquitous use of the internet.  However, once the Commission published its proposal, Mozilla could see that instead of adapting copyright for the next step, the Commission attempted to transplant the antiquated business model of publishers for rightsholders into an online environment.

For her this is especially visible in Article 13 (upload filters), where she identifies four problems. First, Article 13 re-interpreted profoundly the existing liability regime, and removes platforms from the liability exception completely. Secondly, the proposal would require that as a general business model, licenses for content have to negotiated. However, so far, many platforms operated without using licenses, as they are open. Article 13 would change how such companies function. Thirdly, Article 13 required companies online to put in place upload filters to prevent copyrighted material from being spread. Finally, Article 13 did not specify anything regard safeguards, for example what happens when content is taken down erroneously because the content is legal or there is an exception for its use. For MacDonald, the use of filters does not work very well for YouTube, where often legal content is taken off, and there was no reason to believe that they will work well for other companies. Moreover, upload filters are expansive and not every company would be able to afford a tailor-made model, so they would have to rely on the widely available options that are proven to work poorly. MacDonald highlighted that during the legislative process not much progress was made. For her, the European Parliament actually adopted the worst version of Article 13, which would make all hosting platforms liable. Such a version would lead to platforms being overcautious as a sort of damage control, and would result in more content being suppressed. Additionally, the problem of just compensation for artists, as one of the goals of the copyright reform, would not be solved with upload filters either, as it also becomes harder for them to get their work spread and express their creativity.

Regarding Article 11, the press publisher right or ‘link tax’, MacDonald argued that this was an equally bad proposal. The press publisher right was proposed as a neighbouring right and would cover snippets from publisher’s website, so that these publishers can charge platforms, such as Google, for their use. A version of such a right was adopted in Germany, where it has proven to be a bad policy. The big platforms that were supposed to pay now fairly for their use of other companies’ content, have enough economic might to negotiate great reduction or even exemptions from the link tax. Instead of limiting Google it has allowed it to entrench its dominant market position in Germany by taxing only its competitors. A similar proposal also failed to work in Spain, where even publisher were complaining in the end.

Overall, MacDonald summarised, that these provisions will not target big platforms such as Google, Facebook etc. as was aimed for with the proposal, as those are the ones who have the money and economic power to stay in the market. Instead it would hurt their competitors. So far, the only promising development has been in the Council, where they were trying to narrow the scope of Article 11 to make sure insubstantial use (such as tweets) was excluded.

The presentations were followed by an intense debate with the audience, where also a staff member of a Member of the European Parliament gave the latest update on the trilogue negotiations.