In early July 2017 I attended an excellent workshop hosted by researchers at the Centre for Information Technology, Society and Law at the University of Zurich. The objective of the workshop was to bring a group of academic experts together to discuss data ownership rights.
It is perhaps not surprising that the issue of ownership rights in data is bubbling to the surface as we move further into the evolving big data environment. Data have been described as the new “oil” of the information society. They have a tremendous value and are strongly linked to innovation. One of the ways in which industrialized nations have nurtured innovation has been through the creation of intangible property rights such as intellectual property rights. Data ownership rights flow from that same industrial era mind set. However, it is far from clear that this paradigm is a good fit for data and data-related innovation.
The concept of a data ownership right was raised in the EU in the European Data Market Study, Second Interim Report, June 2016. At page 146, it states:
In fact, the way data are made available and the extent to which data are flowing across sectors and organizations, play a fundamental role in sustaining and developing the emergence of a European data-driven economy. In defining and specifying the rights to create, edit, modify, share and restrict access to data, data ownership becomes a pivotal factor affecting a growing number of potential data users and an increasing range of data-related activities.
One might perhaps be forgiven for thinking that there are already data ownership rights; for example, terms of service for websites frequently state that the company behind the website “owns” its data. Canada’s federal government even got its knuckles rapped by the Federal Court of Appeal for making a similar copyright-based claim in one of its data licences (see my post on this decision here). And, while the law of confidential information could be argued to provide a kind of property right in data or information, in reality what is protected by this body of law is the confidentiality of the information. Once confidentiality is lost, it is clear that there is no underlying ‘property’ right in the data.
Policy makers have long been wary of extending IP rights to data – and for some very good reasons. Copyright law, for example, does not protect “facts”, viewing them instead as the building blocks for creativity and expression, and therefore part of the public domain. Of course, copyright law does protect the original selection and arrangement that goes into creating a compilation of facts (i.e. a data set). How extensive this protection ultimately is depends on what a court sees as the taking of a substantial part of that selection or arrangement. It is this protection for compilations of data that no doubt supports those Terms-of-Service claims to ownership of data mentioned above, but the scope and extent of copyright protection in such circumstances is nevertheless limited and uncertain. In the EU, database rights have provided a broader protection for databases, but it still, fundamentally, is not a protection for the data that make up the database.
It is difficult to see where the interest in a data ownership right is coming from. No clear or pressing need to enhance the protection available for data has been identified. Data ownership rights might be more likely to create confusion and uncertainty – and to increase transaction costs and slow innovation – than to improve the current situation. It would be difficult – and hugely problematic – even to begin to try to identify the ‘owners’ of rights in data and to manage the potential competing interests. And while there are undoubtedly issues around the fairness of particular uses of data, or the legitimacy of means used to acquire data, existing laws already offer a range of recourses and remedies that may be applicable in any given case.
The brief summary of our meetings on data ownership is now publicly available, and it addresses these and many other issues relating to data ownership rights. Our conclusions – that there is no evident need for a new data ownership right and that it would be impossibly difficult to define or constrain – offer a caution to those who regard property rights as a panacea in marketplaces of all kinds.