After more than two years of intensive public debate, the negotiations on the European Copyright Reform reached its peak, leading to a compromise agreement among the EU Institutions. The agreed text will now be submitted for vote in the Parliament and the Council.

The text introduces an obligation for online platforms such as Google News or YouTube to remunerate the rightsholders (musicians, writers, performers, etc.). While the legislators agreed that smaller platforms run by SMEs and start-ups need to be exempted from such cumbersome obligations, the agreed exception is way too narrow. Moreover, the proposed legislation creates barriers to European companies, including SMEs, that develop Artificial Intelligence (AI) using machines that perform automatic text and data mining. Rightsholders will be allowed to charge  companies that exploit this technique for using their copyrighted material.

‘Although we welcome the ambitions to rightly remunerate European creators, we regret that a wider SME exception, as proposed in an earlier version, has not been adopted. While the lawmakers aimed at addressing big platforms that extract value from protected content, the agreed text will severely affect SMEs and start-ups.  The narrow SME exception will force most platforms, including small ones, to comply with the same rules as tech giants. Furthermore, authors’ rights on text and data mining will hamper the development of Artificial Intelligence by European companies. This will give a competitive advantage to companies that can develop AI outside the EU, thus leaving European digital SMEs and start-ups behind’- stated Oliver Grün, president of the European DIGITAL SME Alliance. 

Text and data mining (TDM) is known to be one of the main techniques used for machine learning and is crucial for Artificial Intelligence. While European DIGITAL SME Alliance has proposed to exempt SMEs from paying additional licenses for TDM, the final agreement leaves a chance for the rightsholders to decide whether businesses can freely data-mine their work.

Furthermore, DIGITAL SME has stressed that the obligation – to detect and instantly remove unauthorised copies of somebody’s work, is particularly burdensome for small platforms. Having in mind the amounts of data being uploaded every day, the so-called upload filters remain the only way to comply with such requirements. The development of upload filters is not only very costly but also prompted to errors. Therefore, it prevents SMEs from engaging in further endeavours to create new European platforms, and potentially limits the freedom of speech online. What is more, the narrow exemption suggested in the agreed text will only apply to very few and very small platforms. For such companies, scale-up might not be possible at all due to the strict requirements on copyright infringement detection.

However, the agreement reached has also taken into consideration some of the concerns previously raised by DIGITAL SME. To begin with, big online news aggregators and other similar platforms will have to pay their fair share to creators, while they will also have to be transparent about how works and performances of creators are exploited. Secondly, open source development platforms and online encyclopaedia have been exempted from the Copyright Directive, therefore allowing European developers to further evolve free and open software without facing any filtering. Thirdly, parodies, memes and other similar creative expressions do not need to be filtered, thus leaving more space for self-expression online. Finally, snippets of news do not need to be filtered and blocked as long as they are short or have original comments next to them.

Despite the agreement reached, the Copyright Reform is still far from being concluded. The agreed text needs to be approved by the responsible committees in the European Parliament, as well as to pass the final plenary vote and obtain a qualified majority by the Member States in the Council. The following weeks will decide the future of text and data mining and upload filters!