This week I was lucky enough to be invited to say a few words to the Copyright for Knowledge group meeting. I was invited to speak about open science and its implications for copyright and licensing regimes in the UK and Europe but, although I did touch on that, I did rather go off topic to make the point that, unless we think about the policy and regulatory environment more broadly, we will miss important operational and strategic opportunities and risks.
However, let’s start by looking at two recent texts from the European Commission because, while the UK is negotiating its way out of the EU, there are good reasons to think that we will stay closely engaged in the area of research. On 25 April, the EC released two significant texts, one a revised set of recommendations for member states on access to, and preservation of, scientific information, and the other a proposed directive extending the remit of public sector information (PSI) regulations to, among other things, research data.
My colleague Frank Manista reviewed the revised recommendations on scientific information in a post here a few days ago, noting the ways in which they have changed since the 2012 version. Broadly speaking, the Commission has heeded calls for stability in the policy landscape, and has built on, rather than deeply revised, its earlier recommendations. With a new Commission coming next year, following both EU elections and the departure of the UK, it is helpful that this policy agenda seems well-established and bedded in. The proposed PSI directive follows a consultation last year, in which some consideration was given to extending the scope of PSI to cover not only research data but also the administrative information that surrounds academic research, which is held by public research performing organisations. That extension is not in the proposed directive, but research data definitely is, and should be seen in the context of a broader push toward making data available to support Europe’s AI sector. As Natasha Lomas notes in a TechCrunch post, “The changes it’s proposing to the PSI Directive are intended to reduce market entry barriers (especially for SMEs) by lowering charges for the re-use of public sector info; increase the availability of data by bringing new types of public and publicly funded data into the scope of the Directive (specifically the utilities and transport sectors, and research data).”
The focus on data and AI is replicated at national levels, for example in France, President Macron has recently announced a major plan that includes work on both open data and the ethics of AI. In the UK in 2017, the House of Lords outlined some ways forward, and the Industrial Strategy built on this to highlight the potential of AI and data, which has now underpinned a Government-led sector deal on AI that, like the French announcement, stresses both the potential of more open data, and ethical and societal issues.
At the European level, the European Open Science Cloud (EOSC) is the EC’s flagship initiative to support the sharing of data and digital collaboration, and will help the academic research community participate more effectively in the use of AI and other data science techniques. The EOSC Pilot project, in which Jisc is a partner, has recognised that the policy and regulatory environment includes many frameworks, whose scope often intersect or overlap in practice. Examples would include open science policies, IPR, data protection (GDPR), PSI, mandated interoperability frameworks, freedom of information, research integrity and ethics. In January this year, EOSCpilot released a review of many of these frameworks. Its primary purpose was to suggest how they could operate together effectively within the EOSC environment, but its recommendations have wider relevance in the context outlined above. They include recommendations to:
- “Produce consistent policies at the EU, the Member State and the institutional level: The macro policies provide a harmonised framework to a certain degree, but this needs to be complemented by further measures at the micro level. This is a very clear need at the level of IPR policies, Data Protection and Ethics, Procurement policies and Open Science Policies… Monitor the implementation of the EU policies, particularly in areas where frictions may appear (e.g. licensing by Public Sector Organisations, implementation of the General Data Protection Regulation)…. And use fora, along with instruments such as the Open Science Monitor and Registry, to harmonise micro-policies practices, such as licensing of data-sets, access to services through APIs, open access policies etc.
- Standardise interactions at the organisational and institutional (micro) level: Ensure that the Findable, Accessible, Interoperable and Reusable (FAIR) principles are consistently applied at different levels, ensure micro policies are standard, human readable, machine readable and interoperable, ensure the use of open and standard licences or at least standard commercial licensing framework for services consumption or data reuse, ensure the existence of Standard Operational Procedures that interoperate with each other, support massive and semi-automated rights clearance and licence interoperability resolution.
- Automate the application of policies supporting open science by increasing openness and supporting data protection by design and default, as well as data sovereignty for the user. This proposal is the result of the substantial work undertaken in the context of PSI and Data Protection policies. Its underlying assumption is the substantial reduction of transaction costs in the free flow of data. Focus on the semantic representation of licences and policy documents so as to be able to easily search, identify, read and perform compatibility checks between material originating from different sources, use standard model licences and prior informed consent forms…”
It is this kind of quite detailed analysis and review that will continue to be needed as digital research becomes included in ever more policy and regulatory instruments. But also, as some of the points from the EOSC report imply, there is an urgent need for automation here, if the regulatory burden is not to stifle the very innovation it is intended to promote. Part of this automation is through online tools, and EOSCpilot is developing policy recommendations and producing specifications for an Open Science Policy Registry and Monitor which will help to address the identified issues. We expect more detail on these in the next couple of weeks, so worth watching the EOSCpilot website for that.
Open research is best pursued by researchers because it is the best way of doing research; regulation, and associated tools and services, should enable that, but need to be coordinated in order to do so.