Crowdsourcing Human Rights Research with Amnesty’s Decoder Project, Online Discussion Notes


FabRiders’ Network-Centric Resources project helps resource developers understand how to develop people-powered and participatory resources and establish assets for networks and communities that share ownership, enable contribution and support collaboration. In this online discussion, Amnesty International’s Milena Marin shared wisdom and lessons learned about enabling thousands of digital volunteers to participate and contribute to groundbreaking human rights research with their Decoder Project.

There is a vast amount of material online that contain evidence of human rights violations. YouTube alone has over 60,000 years of video stored on its site, with 300 hours getting uploaded every minute. The Decoders Project invites Amnesty supporters to help tackle this by getting them to review online materials and identify if they contain evidence that Amnesty can use in their campaigns for human rights. Anyone with a mobile phone or any other device that can connect to the internet can participate.

The project started in 2016 with Decode Darfur, where over 28,600 individuals contributed more than 9,000 hours to map remote and vulnerable images in Darfur. They recently launched Troll Patrol, asking individuals to join their growing army of digital volunteers to expose abuse silencing women on Twitter.

They have been growing their ranks of digital volunteers by reaching out to Amnesty supporters. If they think a supporter might want to do more than sign an online petition, they will ask them to take part in Decoders. They target specific stakeholders, for example, in Decode Oil Spills, they reached out to Nigerian Amnesty supporters to help identify evidence from the Niger Delta.

The platform is set up so anyone can participate. The site provides would be decoders with a quick tutorial and instructions. They’ve done a lot of work to make sure that people have a smooth experience with as few glitches as possible. A lot of thought has gone into the user design and how people will experience the platform.

Milena sees it getting harder and harder to get people’s attention. Having a strong value proposition is critical. People need to immediately see that their contribution will help in practical and rational ways. Once they have contributed, they need to get immediate feedback. At the end of the project, you need to show how they have contributed. In Decode Oil Spills, decoders were credited for uncovering oil spills that the oil companies never disclosed.

Milena’s process for engagement:

  1. Provide potential decoders a rational and strong value proposition.
  2. Educate participants on the issue and how they are contributing.
  3. Continue to communicate on the progress of the campaign.
  4. Show participants the overall impact of their contribution.

As an innovative project, Decoders has enabled Amnesty to engage more people in its work. There are now more meaningful and diverse ways that people can digitally engage with Amnesty. If you want to innovate, you have to be willing to learn. You have to be ready to hear things about what you are developing that you don’t want to hear. Remember: the people that will give you critical feedback are your real allies, they care.

We discussed the Network-Centric Resources Lifecycle and where Milena sees Decoders within that framework. She hopes the Decoders platform is heading to ‘maturity’ as they have learned so much from each project and applied those learnings. Milena, of course, does not want ever to see Decoders die, but she acknowledges that requires Amnesty to make sure it remains healthy, relevant and useful to people.

We had a great question from our viewers about using open data in human rights campaigns. Milena made the point that while open data is accessible, it’s often not usable. The data collected via the decoders project is as open as it can be: human rights trumps open data. For instance, the Darfur satellite imagery could show perpetrators what villages they missed. It’s easier to make data open in the Troll Patrol project because much of the data is already anonymised. It’s a balance of keeping in mind who you are putting at risk, along with thinking about how can your data support the efforts of your allies.

There is much more innovation on the way for the Decoders platform. They are currently assessing how they can apply all the data they have collected to machine learning. They are now experimenting with developing algorithms that can identify destroyed villages and deploy early warning systems. They are also looking at how to set up algorithms to identify online abuse.

View the full online discussion: