How to make open science work

As open science gains momentum, universities must maintain their academic independence by arming themselves against possible takeover of critical infrastructure, research information and data by private parties.

If universities simply make their data generally available to everyone without any conditions,  commercial entities could collect that data, enrich it and build services around the data, and then make universities pay to use those services.

We subscribe to the idea that collaboration and transparency advance science. We also believe that the results of publicly funded research should be generally available. That goes for research data too. But legal issues and sovereignty issues prevent universities from sharing their data unconditionally.

Legal issues with sharing data

For research data, there is no true ownership, intellectual property, or authorship;  copyright cannot be properly regulated. And the sharing of research data that contains personal information is limited by privacy and informed consent requirements. The General Data Protection Regulation (GDPR) does not provide clarity, especially as it seems to conflict with the EU Directive on Open Data and the Re-Use of Public Sector Information and with Text and Data Mining articles in the EU Directive on Copyright in the Digital Single Market. Data sharing can also be problematic when researchers are bound by embargo periods, limited retention periods or both. Moreover, data use may be restricted by purpose limitations and subsequent bans on dual use and reselling.

Sovereignty issues with sharing data

Data sovereignty means that the data holder should be able to decide which parts of the data to share, with whom, and under which conditions. Data holders are the researchers or institutes that collect and maintain the data. Of course, each university must archive all research data for which it feels responsible. And of course, this must be done in a findable, accessible, interoperable and reusable (FAIR) way; archiving would be useless otherwise. But FAIR does not necessarily mean open. The university may want to impose restrictions on the reuse of the data, or it may even be bound by legal restrictions on sharing the data. Yet, there is another, more important reason to guard data sovereignty.

Commercial lock-ins threaten academic independence

If universities’ data is generally available without any conditions, the private sector is likely to use it to build commercial services that universities will have to pay to use. This threat not only applies to research data and other products of research (software, protocols), but also to universities’ research information systems and funders’ grant management systems. Universities depend on such databases, just as universities depend on scientific journals.

We can’t have our data locked in by private parties in the same way our scientific publications are locked in by commercial publishers. If universities become completely dependent on commercial endeavours, then the independence and scientific integrity of publicly funded research will be jeopardized.

Guiding principles for public-private collaboration

These concerns are widely acknowledged within the academic community. A report by the Scholarly Publishing and Academic Resources Coalition (SPARC)  sets out terms and conditions that must be met for universities to collaborate with private parties: Institutions maintain ownership of data; open procurement of information tools and services; transparency of the algorithms used; portability of the results obtained; no re-sell to third parties; no turnover of sensitive data. A taskforce of Dutch universities, prompted by the recent deal with publisher and data analysis company Elsevier, is drafting a set of principles for responsible management of research information and data, which should serve as guidelines for future negotiations with private parties.

A data market for responsible and secure data sharing

It is possible to exchange research data and other information (e.g. software) in a responsible and secure way that takes all legal and sovereignty issues into account, and that complies with all conditions that universities would like to impose on other parties. The University of Amsterdam is working with the City of Amsterdam, the Amsterdam Economic Board, and SURF (the national cooperative organization of ICT infrastructure for higher education) to create an open, reliable and fair market for exchanging data: the Amsterdam Data Exchange (AMDEX).

AMDEX functions as an interface between existing infrastructures to facilitate a data market in which data holders maintain control over their data. It upholds data sovereignty, separates the data from applications (tools and services), and automates the regulation of data access, the legal recording and enforcement of access conditions. Designed for all private and public markets, including cross-sectorial markets, AMDEX is enabling universities to exchange data between existing data storages, such as university repositories and international disciplinary data platforms.

Enshrining safe data sharing principles in EU policy

The European Commission wants to promote the free flow of data within the EU and across sectors for the benefit of businesses, researchers and public administrations. However, the academic community can only fully participate and contribute to this data-driven society, if our concerns are taken seriously.

First and foremost, EU policy should safeguard the independence and scientific integrity of academic research.  Data sovereignty and principles, such as those formulated by the SPARC and the Dutch taskforce, on the management of research information and conditions for safe data sharing, should be included in EU directives. Second, EU policymakers need to encourage the establishment of a technical interface, such as AMDEX, for data exchange between the numerous repositories, data platforms and storages that are in use by the various data holders. The availability of an interface for secure data exchange would make the European Open Science Cloud (EOSC) a reality in one fell swoop.

Frans Oort is the open science coordinator at the University of Amsterdam. Matthias Bakker, senior policy advisor, and Wouter Los, AMDEX consultant, contributed to this viewpoint.