This is the first of our ‘research diaries’ – where we talk about themes and questions that we come across as we interview people working on human rights documentation as part of our project with Amnesty and Benetech.
We’re sharing this because we want to be open about our process, and to try to involve the many others doing great work in this field. This post introduces a few (but by no means all) of the issues that interviewees discussed.
We’ve talked to five people so far, deliberately trying to include experts with very different types of experience in varying areas of human rights work. While we know we won’t be able to speak to everyone we’d like to, we’re doing our best to take in a range of perspectives across the spectrum of human rights organisations. It’s already helping us to refine our questions and add new ones as we go.
Interviewees generally described new data streams (whether videos or satellite data) as a supplement to more established methods of documentation such as recording eyewitness accounts or victim testimonies. As one interviewee put it, many of the steps for human rights documentation processes have remained the same as they were five years ago – but there’s been a big increase in the types of data available to feed into them (and challenges that accompany this).
As another interviewee highlighted, ‘established’ documentation methods are established for a reason: they work. They went on to suggest that a tendency among technology enthusiasts to describe existing methods as ‘traditional’ can imply that these methods are somehow of secondary importance. We’re keen to bear this in mind throughout the projects, focusing on whether (and how) new types of data can fit into existing processes with well-established standards and proven results.
In line with this, some of our interviewers suggested that bringing in new types of data can often be a lower priority than simply organising and processing the information that organisations already have. Mike Romig, a researcher studying human rights documentation in the Middle East, found that organisations he has talked to are “not really thinking about new data streams: the interest is in going to a more systematised method of documentation as the first step.” Similarly, although SafeCity (an organisation based in India that allows people to report sexual harassment) uses an online platform, Facebook and Twitter to collect reports about sexual harassment, more than 80% of the information they collect comes from on-the-ground work with local organisations in Delhi and Mumbai.
So, how are human rights defenders deciding whether investigating new data sources is worth the effort? Adding new data streams into existing work patterns, by necessity, involves experimentation – and several interviewees described reasons why obstacles to working out whether this is an effective investment of time and resources.
For one thing, organisations can’t always tell at the outset whether using a new type of data will make to a concrete contribution to their work. For example, they might know that they need more information on a particular issue, but is data from social media the best place to find it? It’s difficult for organisations to know what results they’ll get before they have already invested time and resources in testing out a new approach. Moreover, the value of some data streams may not become clear until after the fact: satellite imagery can help to identify things ranging from formal encampments of military units, temporary structures, or movement of communications equipment that may not be immediately essential to a human rights case, but could later provide a piece of information that would otherwise have been impossible to get.
Some interviewees suggested that the more complex the technology is, the harder it is to identify the potential benefits in advance – particularly in relation to techniques like using machine learning to analyse satellite imagery. Several people mentioned citizen journalism and initiatives such as Bellingcat that had clearly demonstrated the value of open source intelligence, combining photos, video and social media updates to bring insights on issues that would otherwise be difficult, if not impossible, to glean
However, one person cautioned that while citizen journalism investigations can lead to a wide range of outputs – many of which can be published quickly – human rights organisations have a much narrower range of potential outputs, and may only be able to show their findings every few months (or even years). Working out the questions organisations ask when they decide to test out a new data source or a particular technology tool is one of the core aims of this project, and we hope that the interviews we still have to come will give us insights that we can share.
Last week, we also hosted the first in our series of community calls on the same topic, where some interesting questions came up, such as:
- how can we best compile, compare and cross-reference older data sets with new/emerging ones?
- how interested are human rights defenders with ‘new’ technologies, versus simply dealing with their regular challenges?
- what are the existing standards and guidelines around human rights documentation, and how do they currently mention technology and data?
To find out more about the call, check out the notes from that call and keep an eye out on this blog for news of the next one. Please tell us if there are topics you’d like to talk about, research or projects you’d like to share, or things that we’re missing. And keep an eye out for the next instalment!