Adapting tech tools for human rights monitoring: lessons from Burundi

A Burundian protester reacts as he is detained by police officers during an anti-government demonstration in Bujumbura, Burundi, 20 May 2015. EPA/DAI KUROKAWA

Evidence of human rights abuses in social and mainstream media has become more and more ubiquitous, and can be used and preserved for human rights investigations, such those taking place in Syria. Anyone with a mobile phone can document abuse: the widespread dissemination of graphic photos and videos during protests against police violence in Nigeria; videos of police and military abuse against protesters in Myanmar; countless images of the assault on the US Capitol. 

Through effective collection, analysis, and strategic use, such abuses can—and should—be reported to international human rights bodies. As illustrated by an initiative in Burundi, which was kept confidential until now for security reasons, technology can help evidence collection reach the threshold required by international human rights bodies. 

In 2015, when Pierre Nkurunziza of Burundi announced that he was running for an unconstitutional third presidential term, Burundians took to the streets. The Burundian government was responsible for massive and widespread repression against protestors, including a rapidly suppressed attempted coup. Arbitrary arrests and detention, disappearances, torture, rape and other forms of sexual violence were among the violations against any and all forms of perceived or alleged opposition. Human rights defenders were hastily forced to leave their lives behind.

In this context of acute crisis, a number of initiatives to document the considerable scale of human rights violations sprung across the country, often with international support. Courageous rights defenders and many ordinary citizens started putting their witness to paper or simply took pictures and videos, sometimes on social media and often at very high risk. The US-based Carter Center and the Centre for Civil and Political Rights launched a “Technical Assistance Project” (TAP) initiative that repurposed a digital tool for election monitoring in order to better document such abuse. 

Courageous rights defenders and many ordinary citizens started putting their witness to paper or simply took pictures and videos, sometimes on social media and often at very high risk.

Because of the highly volatile situation in Burundi back then, this initiative had to be confidential: the risks to those involved on the ground were significant. The tool enabled the broad collection of direct witness reporting in a systematic way. Cases documented through the initiative led to, among others, a confidential report submitted to an ongoing preliminary investigation by the International Criminal Court which registered it in July 2016. 

The report, which provided evidence of daunting violations, was also submitted simultaneously to the UN Committee against Torture ahead of a rare and special review requested by the Committee. In addition, there was a public joint NGO report, which the Burundi government delegation said they “were not able to consult with enough anticipation” as its justification for missing the second part of the review—an unprecedented affront to this UN body. 

Well-documented reports such as the one provided through TAP aided in the international recognition of massive human rights violations in Burundi. The UN Human Rights Council went on to establish a Commission of inquiry only one month after the special review by the Committee against Torture, through a resolution which cited Burundi’s lack of cooperation with UN mechanisms. The Commission continues to operate to this day. 

For its part, the International Criminal Court subsequently opened a full investigation in October 2017, raising hopes for an end to impunity in Burundi, despite the government’s decision to withdraw from the Rome statute. 

Thankfully the human rights situation in Burundi has slightly improved (as illustrated by the recent release of defender Germain Rukuki), due to a range of factors as well as pressure from international human rights instances. Nonetheless, much work remains: human rights activists continue to be judicially harassed, and the international Commission of inquiry still has no access to Burundi. 

The tool used in this context (NEMO, formerly “ELMO”) combined features of cutting-edge crowdsourcing platforms such as Ushahidi or the newer Uwazi, with features tailored to meet the security and documentation demands of human rights reporting in a highly repressive environment. Several adaptations and considerations made NEMO a good fit, of which we highlight three:

First, flexible technological input. NEMO enabled data collection via mobile devices, desktop computers, and even SMS. NEMO also did not require constant connection in order to collect and transmit information. This meant that in areas with low bandwidth or no data connections, monitors could still collect material for reports on their devices to transmit at a later time.

Second, real-time multi-lingual reporting and analysis: NEMO’s templates enabled users to submit data in their native Kirundi or French, which was subsequently analyzed in the original language as well as by English speakers. Because the system could process submissions in multiple languages, team members had access to real-time analysis in the language of their choice regarding categorical questions about region or types of incidents.

And third, security & encryption: Human rights monitors put their lives at risk to defend the rights and freedoms of others. Designing a platform that gave them the best protection while still using cost-effective open-source technology was a necessity. We worked with the Georgia Institute of Technology Research Institute who penetration tested and reviewed code to help ensure that NEMO had robust measures against hacking. Because open source technologies are often built with components of varying quality, we wanted to assure our users that they were using communication technologies built according to the highest possible security standards. 

Some of the challenges that we encountered included maintaining the anonymity and confidentiality of users on the ground; learning how to use the tool; building skills required to document human rights violations; and ensuring the safety of monitors on- and off-line. Not every monitor was able to remain with the project due to challenges that were unavoidable. 

From our experience we learned how to design a mission to consider multiple data inputs, since people are collecting that data in multiple ways (including paper); design it for multiple languages (or even more colloquial language that can be easy for general citizens to use) and design it from the outset to high security standards and training, which takes time. The more complicated a technological system is, the more opportunities for security failure. Contexts may require certain parameters—transmitting coded information via SMS, such as was implemented in a version of NEMO after Burundi was more optimal for network capacity in the Democratic Republic of Congo but it also meant that less narrative reporting was possible; it also required additional training with regards to security in case coded SMSs were intercepted, but the risks were lesser. Designing and implementing technologies according to a human right perspective is an ever emerging field, so we look forward to seeing how other human rights defenders and supporters continue to advance reporting capacities in the years to come.


The authors would like to thank Gabrielle Bardall, The Carter Center, and Friedhelm Weinburg for their feedback on this post.