In August I participated at the fifth annual conference of Hedayah, the International Center of Excellence for Countering Violent Extremism (CVE) which took place in Swansea, UK. Experts, researchers and policymakers present discussed a range of topics such as the continuing threat of Daesh, the extent to which the far right presents a democratic challenge, the role of gender in radicalisation, and the application of reintegration and rehabilitation methods.
I spoke and led a breakout session on ethical research in the field of extremism. Drawing mostly on our recent project and my experience in the Western Balkans, I focused on ethical research in online radicalisation. I discussed general concerns for 1) the ‘do no harm principle’; 2) privacy and anonymity issues; 3) online consent; and 4) data reporting.
The applicability of the ‘do no harm’ principle in online research is very complex and needs a thorough consideration. This is not only vis-à-vis online users whose user handles may be incorrectly ‘labelled’ in such research as pertaining to radical categories but also vis-à-vis researchers who leave digital footprint. During the presentation, I also discussed the problem of ‘false positives’ in machine learning in online radicalisation, i.e. the incorrect labelling of content and users online via large data-collection methods.
Privacy concerns in online research are equally important given the increasing scrutiny over data anonymisation and protection of online identities especially in the European context. While research in the United States pertains to different concerns for data privacy – as such research is often justified as ‘public good‘ for security reasons – European organisations need to abide to a higher set of privacy standards introduced by the GDPR recently. This poses a number of dilemmas for EU-based organisations.
Online consent was another point of discussion. As users online are often not aware that they are being observed, direct or implicit consent is hard to obtain without compromising the research or researcher. I offered several examples of strategies how to mitigate this challenge by discussing the concept of ‘public data’ and how that affects the need for online consent.
I concluded my session noting that everyone conducting research on extremism and radicalisation must be able to report the collected data ethically (i.e. anonymising where necessary, not distorting the data and ensuring no participant has been endangered by the research) and draw up a set of ethical red lines that must be inherent in all project implementation and data reporting.