by Jan Willem Knibbe
Text and data mining is a hot topic. Some see it as a new revenue source for market research and insights professionals, with its promise of effective analysis of vast amounts of data. You don’t have to ask respondents to fill in cumbersome questionnaires but you can analyse what they post online. Fast, cheap and versatile data. However, in Europe the current copyright framework poses a big hurdle for these types of research projects.
Under the current legislation, researchers interested in data mining can only copy copyrighted data with the permission of the author. Unlike in the USA, there is no fair-use provision in the legal framework that would enable using these data for research purposes without the explicit permission of the author. This gap has also been identified by the European Union. As part of the Digital Single Market agenda, a review of the copyright legislation is foreseen. One of the aspects that this review should cover is creating an exemption for research purposes in the copyright law. We have previously published an article about these developments: A new copyright directive in Europe: will there be room for data mining?
Since we published that article it has now become clear that the European Commission will champion a research exemption in the new framework. However, a debate has emerged whether only non-commercial research should benefit from the exemption. This is a revival of a notion first proposed in the General Data Protection Regulation where it was first suggested that only non-commercial research should be allowed to make use of any derogations foreseen for research purposes.
This position, however, is not unanimously supported. Support for broadening the exemption to all research isn’t coming just from business research community, but also the academic research community which issued a clear call that any distinction between commercial or non-commercial research is artificial and will only lead to legal uncertainty. For example, LERU (League of European Research Universities) writes that the potential of Text and Data Mining (TDM) is acknowledged by researchers, who see the benefits of using automated tools to mine the literature and supporting research data. (…) However, the legal basis to allow the use of TDM techniques, certainly in licensed commercial literature, is unclear. What is needed at a European level is a Fair Dealing Exception, certainly for the purposes of research, in the EU Copyright and Database Directives to facilitate the sharing and re-use of research data.
One of the arguments in favour of not making a distinction is that it would be legally difficult to determine where to draw the line between commercial and non-commercial purposes. It has been noted by many that commercial research projects are done by universities, while commercial companies are executing research project for the public good. This has been partly recognised by DG Connect by focusing on the purpose of the project, and not the body executing the project. Nevertheless, the question on where to draw the line remains open for interpretation and consequently which projects might benefit from the Commission proposals.
Another issue being addressed it the current voluntary nature of any exemption in the Infosoc Directive, which is the European legislation governing copyright. This means that every Member State is free to choose whether or not it provides for an exemption, including for text and data mining. This creates a complex, heterogonous legal framework to contend with, especially in the case of international research projects not to mention the uneven playing field created in the different Member States.
Unlike the discussion around putting limits on research, this seems to be a more accepted change in the legislation. Nevertheless, European harmonisation is increasingly sensitive in most national governments, who have to approve any new directive in the Council of the European Union. It remains to be seen if this proposal will make it to the finishing line. A new legislative proposal covering data mining is expected in September 2016, but it yet unclear when it could be adopted by the Parliament and Council. For an explanation of the process, please consult our law making in Europe beginners’ guide.
All in all, there is still a long road ahead of us. ESOMAR will continue to monitor this dossier, and try to convince European legislators that the distinction between commercial and non-commercial research is artificial would prevent all research actors, including applied research actors like market research agencies, from using text and data mining to improve the speed, relevance, and effectiveness of insights for the betterment of businesses and society alike.
Jan Willem Knibbe is Policy & Public Affairs Assistant at ESOMAR.