Editor’s note: This post forms part of a larger series addressing key issues related to human rights, technology & big data.
New data streams present great opportunities. They are revolutionising how we work every day. Improved Internet and mobile phone connections have changed how we order a taxi, how we book a hotel, how we watch television. It has also changed opportunities in how we communicate with the world. Scenes of nearly every global event are written about, filmed or photographed thanks to the smartphone in the pocket of a witness to events and posted to a social media site. From sports events to elections, from celebrations to disasters. News reports and images from every news event now incorporate some element of social media – be these the pictures that launch the story (think of the shooting of Alton Sterling by Baton Rouge police in the United States in July 2016), be these the first pictures of a major news story (the terrorist attack in Nice, France or the attempted coup in Turkey) or reactions to ongoing news stories.
While social media has revolutionised news gathering and storytelling, it has also opened enormous possibilities for monitoring human rights violations and humanitarian crises. In January 2016, Amnesty International launched a report into mass graves found in Burundi based on satellite imagery backed up with social media eyewitness reports on the ground and traditional interviews. Human Rights Watch has used content sourced from social media to highlight potential human rights abuses around the shooting of Shaimaa al-Sabbagh in January 2015. WITNESS trained and encouraged activists in Rio de Janeiro to film potential human rights abuses in the run up to the summer Olympics in Brazil’s second largest city.
In the humanitarian space, much is also being done. The Humanitarian Open Street Map Team uses open source and open data for humanitarian responses to earthquakes, such as the one that hit Ecuador in April 2016. The purpose of this project is to create, online, the maps that enable responders to reach those in need after major disasters strike in the world. The Open Data Kit helps organizations collect mobile data. These tools have been put into use after the Nepal earthquake and, more recently, by the Bolivian Red Cross to combat the Zika virus.
The possibilities of using these tools to advance human rights and improve humanitarian work is the focus of the Human Rights, Big Data and Technology (‘HRBDT’) Project’s fourth Work Stream: Advancing Human Rights and Humanitarian Responses. And the possibilities, on the surface, do seem endless. But are they truly? Can all this data that is collectable from the public, from eyewitnesses in crises, be useful to prevent human rights abuses or to help those in an unavoidable crisis? And, if they are helpful, what kind of tools really do help? While the appeal of big data and its uses for humanitarian work is great, there is much scepticism. As Sandvik (2016) notes: “technology alone cannot fix broader, long-standing problems, such as ineffectual information management, which can lead good data to go unused. Similarly, as evidenced by the 2011 famine in the Horn of Africa, comprehensive early warnings do not necessarily translate into effective political action.”
Understanding and addressing such concerns are key to all of the different sub-streams of work that will be happening under Work Stream Four, and indeed the HRDBT Project as a whole. We are asking what projects already exist, what have they achieved and how can they be useful. We are working closely with organisations – intergovernmental and non-governmental – that can use data gathered from social media to monitor humanitarian situations to see how we can be useful to their work, and we’ll be building tools to help.
First, we’ll be building tools using text-based social media sources. To do this, we will conduct a retrospective case study of a recent event involving a humanitarian crisis where we know a large volume of social media text was available. The aim here is to find features that might be helpful to automatically identify emerging humanitarian crises by developing computational “text analytics” methods to identify patterns of behaviour on social media, and other sources. The final step of this sub-stream of Work Stream Four will be to evaluate these techniques using the data collected as part of the case studies, to identify, and ideally give early alerts of potential humanitarian crisis.
Second, Work Stream Four will also examine methods for making sense of visual content shared on social media. This can be eyewitness videos and photographs and other sources such as satellite imagery. Similar to the above, the objectives will be to automatically identify patterns of behaviour that are indicative of humanitarian crisis and to assess how humanitarian organisations can strengthen their use of eyewitness media. The case study will apply existing computer vision algorithms, as well as develop new ones. The goal will be to detect similarities in multiple images and videos, to determine when two or more videos capture the same event from different perspectives, and to detect objects such as weapons. The overarching aim here is to help with the verification of imagery to ensure hoaxes or misleading information is not being spread via social media and to try and gather information that can help those organisations whose mission is to help in humanitarian crises – such as, for instance, the movement of refugees.
Third, in collaboration with one of our Project partners, the Human Rights Data Analysis Group, Work Stream Four will assess the extent to which the use of big data, in particular social media, can enable the accurate estimation of the number of human rights violations. This determination of the extent to which it is possible to get the numbers right will occur through a number of ongoing and new projects, including a study on the number of victims in the Syrian conflict killed in detention facilities.
As we move on over the coming months, we’ll be speaking to non-governmental and intergovernmental organisations to ensure we’re on the right track and building tools that help the work of humanitarian organisations in a sustainable manner rather than hinder. Do get in touch with us if you’d like to collaborate with us or learn more.
Disclaimer: The views expressed herein are the author(s) alone.