Using satellite images and volunteer labor to identify bomb craters. Photo from a session at MozFest 2015.

Finding Alternatives to “Clicktivism”

This is a transcribed and edited interview with Milena Marin who runs Alt-Click at the Amnesty International. It was recorded for the NetPosi podcast at the International Secretariat in London, England. You can listen to full audio here.

Drew: Tell me a little bit about your work, Milena.

Milena: I run this project called Alt-Click, which stands for Alternatives to “Clicktivism”. I’m exploring various technologies that we can use to engage people in documenting human rights abuses.

We have all sorts evidences of human rights abuses all over the world. Our researchers work really hard to mine this evidence and really expose and shed light on human rights abuses. Instead of having only researchers drive this, we are really keen to engage everyday people from around the world in this process.

Just to give you an example, we very often work with satellite images, especially in places that we can’t reach because they are closed or dangerous. The researchers can’t go on the ground. We use these satellite images to look for things like military presence, to look for building destructions, nature degradation which can be sign of exploitation of natural resources and things like this.

We use analysts at the moment to analyze images and to look for specific signs of the things I was mentioning before. Instead of doing that, what we plan to do is open up these satellite images to people around the world and have them help us look for these signs. We think that we can be faster, more efficient, and we think that people can have this exposure to what human rights abuses really mean and how they can get into it.

Drew: So, tell me a little bit more about that. So, an analyst will look at satellite images. What will they be looking for?

Milena: For example we work quite a lot with illegal demolitions and forced evictions. There are people who live in slums all other the world and governments actively try to get people out of slums. They do so through illegal means, very often. They would just demolish houses and force people to leave their homes and so on. You can observe that from satellite images quite easily. You can look at the satellite image before the demolition and after the demolition. You can actually see the impact and the scale in numbers. You can see how many buildings were destroyed, just as much as you can see how many trees in a forest. And you can see how many people died as a result.

We look for things like this in war situations, heavy military equipments, craters, or scars that are left by the bombings. We look for pathways that, for example, tanks used to get into a certain city. We work as well on exploitation of natural resources. We look, as I said before, for dead nature. You can quite easily see alive versus dead trees in satellite images, using different light spectrums. We look for that as an example of natural pollution. We can see for example there was oil expedition and pollution because there’s a lot of dead nature around an area. So, these are the things we look for in satellite images.

Drew: So, right now at Amnesty, there is staff people who do this internally?

Milena: Yeah, we have analysts and researchers that do this kind of stuff.

Drew: So, your idea with alt click to open that up and have different people working on it too?

Milena: Exactly. We use a technology called “micro-tasking” that allows us to take a fairly large satellite image and break it down in small parcels and have people analyze each parcel, one-at-a-time. Then we would have a puzzle of many kinds of contributions from all over the world that we piece together and reconstruct that satellite image. So we take the small pieces and put them together and see how it looks.

Drew: That’s cool. How’s that done?

Milena: The technology that exists, it’s called micro-tasking. You might be familiar with it from a commercial service called Mechanical Turk from Amazon. This uses a contributions from humans around the world to do tasks that typically need that human contribution, so they need a person to look at it. You can’t just program an algorithm to look at it. It exists, it’s just about being able to apply that technology to the sphere of human rights.

We have different types of data that we’ll be looking at, satellite images are one example but we look at documents, videos, images, social media reports. We have different types of actions that people can do through micro-tasking. In the satellite image, people could count buildings, count tents in a refugee camp, count the number of tanks in a given neighborhood and so on. They could also outline or trace. So for example, we are looking at the shape of buildings. If we want to look at how prisons look from above. Some of them are bigger, some are smaller. Some of them have different infrastructures. If we look at that, we can ask people to trace or outline shapes of things. We could ask people to identify specific features like in insignia. For example, if we want to understand which military division was monitoring a protest, we looking for specific insignia on military or police to see what kind of division was there. All sorts of things like this are possible through micro-tasking — it’s taking a fairly large and complex research task, being able to break it down in small pieces, that people, individuals, can help with without much training. We don’t require people to be experts in satellite images or experts human rights. They just have to have an Internet connection and be able to have some time to dedicate to us and be a little bit interested in human rights work.

Drew: Tell me more about these people. Are they mostly supporters? How do you connect with them?

Milena: Hopefully they are supporters, but hopefully we can reach new audiences. Amnesty International is funded by people around the world. Ninety five percent of the our funding comes from people. They are generally interested in what we do. You get involved with our work beyond just giving us money. They write letters for people that are facing terrible situations like death penalty or really terrible abuses of their rights. They sign petitions in support of our campaigns but we want to give them different ways, more meaningful ways, to engage with our work.

Hopefully we can reach existing supporters and know more or less who they are, but also we think that we can tap into new audiences. Maybe young people who have come out of universities or people that are looking for meaningful interesting ways to spend their time. Hopefully we’ll tap into both, new and old.

Drew: What do you see is the outcomes of someone running a micro-tasking project for Amnesty? What would be good outcomes?

Milena: For Amnesty International or for people who engage with us?

Drew: Actually, for either. Are they different?

Milena: Well, I think so. For us, we have to have two big goals. One of the goals is to be able to analyze that data that documents human rights abuses — to be able to be faster, quicker, more efficient, and analyze those satellite images that I was talking about, and to process the huge amount of footage coming out of war situations. So having those contributions from other people is really valuable for us. That is an important outcome. Also the engagement and being able to offer people different avenues to be part of our work. While I think when people engage with us, it’s interesting to be exposed to the work we do in a different way. Maybe they’ve heard of Amnesty, they’ve signed a petition, they’ve read the report and they know that terrible things are happening around the world but being able to engage directly with them and having that experience in analyzing, first-hand, the evidence of human rights abuses is very powerful. People get very empathetic with the issue. I think that’s very interesting. As we get more advanced and be able to involve more people, I think people will be able to gain interesting skills.

One challenge we have is verifying video content. We have lots of videos from conflict situations. If you just think of the amount of videos that are coming out of Syria, it’s almost impossible for us to be able to verify those and be able to actually use them in our research. But if we had more capacity and if we had more people that are trained and skilled at verifying videos, establishing where — geolocating videos using open source tools or using Google Earth or different kind of tools, to being able to time a video, to being able to verify the source, and so on. There are interesting techniques that people can learn by working with us. So hopefully, they’ll be able to offer that as well.

Drew: What do you see as barriers to this? What do you see as the stumbling blocks for a project like this?

Milena: Well I think one of the biggest barriers is around verification. If we ask people to help us document human rights abuses or help us understand what is in a satellite image, our researches are quite skeptical about contributions from people around the world. We typically work with people who have two PhD’s and are very experienced in research and have solid contacts on the ground and spend a lot of time on these issues. They would be probably quite skeptical about data that is coming from 2,000 people from around the world. And very likely we’ll have malicious people that will use these tools just to skew our results. Depending on the contexts, people have different views, and it’s really hard to get them to agree on something. I’m sure if we work on Syria, for example, there will be people that will use these tools just to demonstrate a point. So, verification would be quite a big barrier for us.

Drew: Do you have any ideas of how to mitigate that?

Milena: Well, yes, I’m sure it is not an insurmountable barrier. There are verification techniques built into the micro-tasking tools. If we’re using satellite image and we’re asking people to tag tanks in the satellite image, we can set a threshold so that only if10 people agree that that is a tank. The tool will give us a point of data. So, if there’s a disagreement, if 5 people say it’s a tank and 5 people say it’s not a tank, maybe a researcher can look at it. There are mechanisms within the technology to establish or to gain a consensus about something. Then, I think, our researchers would have to look at some of the content to verify. They’d had to use their judgment to see whether it makes sense or it doesn’t make sense. It’s doable and it will be extraordinary to have that contribution to people. This is a small challenge compared to what we gain. We gain having lots of people engaged in our work, them having a new an insight into how we expose human rights abuses, and into that new mechanisms, and into the evidence we have and how we analyze it. I think it’s extremely valuable to be able to offer that.

Drew: If everything goes really well, what are your hopes and dreams for this project? What are the big visions for microtasking and outclick at Amnesty?

Milena: My hope is that, first of all, researchers and campaigners at Amnesty International have the means — meaning tools and techniques — to engage lots of people from around the world in their work, to really expose the in’s and out’s of their work and be able to welcome other people to contribute. That can be micro-tasking but would hopefully would go beyond that, hopefully, we’ll be able to expand beyond to other tools, techniques with offline and online approaches. We start with micro-tasking because it’s probably the easiest one, it’s the low-hanging fruit, although it’s quite complicated. But, yeah, my hope is that we would be able to engage people. We would be able to expose our work and welcome others to look under the hood of the car.

Drew: Cool! So any last thoughts or shout outs for someone who might be interested in getting into this kind of work or learning more about what your working on?

Milena: I would for people to help me as soon as we have pilots. Now, we’re at the stage where we are preparing some pilots and once that it is ready, I would to have people to provide us with some thoughtful feedback about what works, and what doesn’t work, and what would incentivize them to contribute more. So I would say keep an eye on the Amnesty.org, follow me Twitter (@milena_iul), ping me. I would love for people to share their thoughts with me. I would love to be able to share tools whenever we have them.

This is interview is originally from NetPosi, a podcast about activism and technology. You can listen to the full interview and you if you want to hear more interviews like this you can subscribe using iTunes, Soundcloud,or by adding your email on netposi.com