Introduction
Part of the basic objectives of ROSiE are to guide and equip open science (OS) users after exploring the ethical, integrity, social, and legal challenges of OS; mapping existing technologies that safeguard responsible OS, and engaging stakeholders in an open dialogue. Science relies largely and increasingly, especially for its opening process, on research infrastructures which are either (i) generic in their scope but dedicated to OS such as OpenAIRE, D4Science, etc., or (ii) thematically focused but with a commitment to opening scientific processes and especially the release of data and publications such as DARIAH-EU, BBMRI-ERIC and other ERICs (European Research Infrastructure Consortia). Most EC-supported research infrastructures have a solid commitment and experience in addressing responsible OS issues. ROSiE sees the analysis of these infrastructures’ current practices as a robust first step towards its goal. For this purpose, a workshop was organized online on July 2021. The workshop was co-branded with EOSC and it gathered a large panel of ERICs covering a diversity of scientific fields. The participants were specifically invited to witness and comment about (1) the RE/RI/legal challenges and gaps they identified in their respective fields, (2) the (types of) solutions they have implemented so far and (3) their expectations of ROSiE and the form ROSiE’s outputs are to be funneled through the Knowledge Hub, the guidelines, the trainings, etc.
General considerations and limitations
As expected, the overall organization of the discussions according to the status of each issue (challenges, current implementation, expectations) did not favor a clear distinction to be made between research ethics, research integrity, and the social and legal aspects of responsible science. Instead, most of the attention quickly became mainly focused on two matters, (1) compliance to the GDPR and (2) informed consent.
It was also observed that much of the focus was on data during the workshop: this is only one of the products of science, but an illustrative one.
One participant suggested that Gaia-X be considered close to ROSiE in its objectives. Gaia-X aims at creating “the next generation of data infrastructure for Europe, its states, its companies and its citizens”, that is an ecosystem ensuring digital sovereignty and fostering innovation based on trust in the digital environment. Indeed, ROSiE expects its outcomes on the determinants and practical tools ensuring responsible open science in Europe to be useful for such a strategic endeavor as Gaia-X’s.
Design and enforcement of policies, principles and norms
One of the main challenges mentioned by the participants during the ROSiE-EOSC workshop is that ERICs in particular, and other research infrastructures as well, feel that they have little or no power to force users to open up their collected data, and even more so in a responsible manner. Rather, this power lies in the hands of the institutions that employ researchers, and/or can only be enforced by national policies in each country.
However, the participants agreed that research infrastructures are in a position to raise awareness, to facilitate and encourage data sharing, to offer training, and to elaborate guidelines and practical tools for this purpose. Because they are by construction a meeting point for coherent communities from all origins, research infrastructures have the ability and the legitimacy to structure their respective communities and their sets of practices. This is an especially useful external resource for researchers in smaller universities and structures which do not have enough in-house resources such as the legal and/or ethics specialists employed in dedicated offices of larger universities.
Guidelines and training were believed to be essential, but they must be discipline-specific and/or area-specific. Research infrastructures, often being discipline-specific themselves, are therefore in a good position to implement or disseminate them, to stimulate or even participate in their design, to raise awareness among users and all stakeholders, etc.
Altogether, thus, the layers of (1) setting up a policy and general principles, (2) translating these into workable guidelines, norms and tools and (3) applying these in daily practice are operated by different types of stakeholders. These stakeholders interact and co-construct at all layers but the national and institutional levels are preeminent in (1), individual researchers and end-users in (3), and disciplinary or otherwise field-specific interfaces such as research infrastructures are in an excellent position for (2).
Examples of current implementation
Some ERICs maintain legal information platforms for their communities. Examples are:
- BBMRI in the field of biology/biobanking through its ELSI – Ethical, Legal and Societal Issues services.
- CLARIN in the field of social sciences and humanities through its legal information platform.
However, the research infrastructures reported difficulties to keep such platforms updated, especially with the diversity of member states’ legal regulations of intellectual property. The ELSI services of BBMRI have published their return on experience.
An example of how institutions deliver guidance on their policies is provided by Delft University of Technology (NL). The general principles guiding the staff in all processes of their academic practice are listed, however without much details since those are better established on a smaller scale (i.e., disciplinary).
Funding agencies can also provide such information services, like for instance the platform on research integrity maintained by DFG, the main German research funding agency.
Depending on the universities and institutions, the support for the implementation of responsible science is provided by one or another office. All aspects of responsible science may not be addressed by a unique office in a given institution. For instance, at Stockholm University, while patent and intellectual property support services are often provided by the legal office, research ethics and research integrity services are provided by REIS, the office that provides services on research funding matters and the office which researchers are more familiar with.
Remark: maintaining specialized services is costly
When research infrastructures function as providers of tools to guide users through legal aspects, especially those linked with GDPR, they stress the elevated maintenance cost of such services. Moreover, the challenges faced for the maintenance of such tools are numerous, as exemplified by ELDAH CFW.
The challenges for maintaining specialized services are not only in terms of costs, but also in terms of human resources, i.e., legal or other specialized competencies are challenging to hire. Their occasional turn-over, for instance when a specialist leaves, is also difficult and comes along with their successor not feeling comfortable updating and amending the work done earlier by someone else. Moreover, legal offices in universities and research institutions are understaffed, taking into account the growing importance of GDPR in research. Finally, such platforms target an audience that is not accustomed to legal questions. The tool developed by legal specialists has to be understood by a non-specialized audience, making set-up and maintenance even more complex.
Informed consent: some clarification is needed
In terms of informed consent, it was pointed out by the participants that it is sometimes difficult for researchers to distinguish between (1) the legal basis for the processing of personal data and (2) the fundamental ethical requirement for research involving human subjects or participants. The need for education and user-friendly instruments is evident, or the collection of informed consent may be seen as undue bureaucracy by researchers, with the risk that it is eventually overlooked or addressed aminima.
It was also recognized rather unanimously that signing a paper or ticking a box on an online form is not a sufficiently robust expression of informed consent (or consent expression at large), despite the sometimes-different traditions in different EU member states. It has to be the result of a more dynamic interaction with the community. Some institutions, especially Scandinavian ones, such as, e.g., the University of Stockholm (Sweden), officially advise researchers not to use informed consent as a legal basis for personal data processing even if, of course, they still need to obtain consent for participation in a given project or cohort.
Two additional specific challenges were identified:
- Childrens’ digital rights.
- In the Humanities, where primary data is often attached to actual persons, intellectual property is often managed as copyrights by scientific publishers who only rarely approve of delivering them to re-users.
Examples of current implementation
An online consent form wizard (CFW) has been co-developed by DARIAH (ELDAH – Ethics and Legality in Digital Arts and Humanities) and CLARIN in order to support researchers in the field of Humanities in the process of collecting valid consent for data processing in the context of their specific professional activity.
The project PANELFIT has recently delivered an analysis of the issues and gaps regarding informed consent in the domain of ICTs (information and communication technologies).
A geographic heterogeneity of practices
The different national traditions and practices between EU member states sometimesprevent ERICs and European infrastructures from establishing common guidelines at the European level.
For instance, while the University of Stockholm advises against using informed consent as a legal basis for further data processing and/or re-use (see above), which fits with, e.g., Scandinavian traditions and practices, for historical and cultural reasons Germanic countries such as Austria and Germany put emphasis on actually doing so.
In general, it may be difficult to apply common guidelines uniformly: rather, in some instances they must be adapted to national specificities. ERICs and other research infrastructures, therefore, must aim for flexibility and compatibility between countries.
Practical implementation of GDPR was also reported to differ between member states.
Examples of current implementation
The section (see above) about the successive layers of policy/implementation/practice applies to this kind of geographic heterogeneity.
Earlier in 2021, EOSC-Pillar has delivered a documentation of the legal and organizational aspects of service delivery for open science in Europe. Much of the attention regarding responsible science as defined here is given to legal and ethical aspects (especially but not only GDPR) while research integrity and social aspects remain to be explored, although the 2019 brief published by SPARC Europe on “RI through open science and FAIR data” is referred to.
Not all science products can be responsibly opened
Another challenge that arose in the discussion was linked to business confidentiality and the way it can be conciliated with openness and data sharing. For infrastructures involved in partnerships with industrial or private bodies, this requires discussions and agreement signing on this topic before the project starts. At this point, it was reminded that FAIR data does not necessarily mean open data. It is possible, according to the involvement of academic or industrial partners on a topic, to adapt the data sharing rules of the consortium. Therefore, the guidelines must include conditions and possible restrictions for openness.
Examples of current implementation
The FAIR principles published by Holub et al. (2018),[1] provide a shared guidance on what products of science to open, and how.
In the digital era, a diversity of publications needs to be qualified
Information and communication technologies have a profound impact in the way all steps of research are performed. Anyone can publish anything on social media or on more formalized or specialized platforms. While opening science is definitely a positive endeavor, there is a need to qualify these publications and to audit their results, following processes that will probably be as diverse as the form they take themselves.
Remark: relying on data management plans (DMPs)?
It was proposed by some participants that DMPs could be considered as a basis to implement some guidelines (non-discipline specific). A DMP could be used as a way of mapping to provide procedures and guidelines and translated into a roadmap and/or a tool of self-training for researchers. Such early mapping of the project data and their attached challenges in terms of responsible science would help structure the projects at an early stage, anticipation being considered a key to handling legal and other responsible science issues.
Notes
[1] Holub, P. (2018). Enhancing Reuse of Data and Biological Material in Medical Research: From FAIR to FAIR-Health. Biopreservation and Biobanking, 16(2): 97-105. doi:10.1089/bio.2017.011097
This passage is part of D6.1: Preliminary analysis and mapping of existing European and national Open Science infrastructures with regard to promoting responsible Open Science