February 23, 2022, the European Commission has released its awaited new proposal for the Data Act. Among the aspects the Data Act covers is  the clarification of certain aspects of the Database Directive. Technopolis Group, together with Copenhagen Economics, CSIL, and legal experts from Tilburg University (TILT) and the University of Göttingen provided the support study for the review and impact assessment of the Database Directive for the European Commission, DG CNECT.


The Database Directive (Directive 96/9/EC), established more than 25 years ago protects the intellectual creativity embodies in the design of databases. It is an important piece of legislation for example for commercial database providers as well as for academic researchers. Following an evaluation in 2018, the Directive was however characterised as not being fit for purpose in the new data economy, primarily due to the unclarity of the sui generis right in relation to machine generated data (MGD). In the area of internet of things (IoT), robotics, smart devices etc. provide through embedded sensors terabytes of data that could be used for new data driven business models but also to enhance competition and lower prices for end-users if appropriate access rights can be agreed upon. Yet, a number of legal issues and opposing stakeholder views hamper a thorough revision of the Directive. Among the key factors is the lack of legal definitions: what is machine generated data? Is it created or obtained? Can machines create? The latter question sounds familiar? In fact, legal academics and policy makers alike struggle to provide legislation on AI and defining the scope of ‘creation’. Existing case law is so far insufficient to provide clear guidance.

Based on a stakeholder survey and interviews with legal academics, the views on the legal options for including MGD in the scope of the Directive were divers. What seemed to be very positive for some industries seemed to pose threats for others.

Very concerned were also academic researchers, fearing that a lot of data would end within closed databases or at least with difficult and costly access models.

The EC has taken the collected evidence and decided against a major revision of the Directive. A future-proof definition of ‘machine generated data’ is currently out of reach.


The study used primarily qualitative means such as a legal literature review, a survey, as well as internal and external discussions and interviews with legal experts to reflect on various options. Since the study followed the Better Regulation guidelines, the costs and benefits of options had to be estimated. Relevant data was collected through the survey.

The study can be found here.

What's new?

All articles All news