Visual media technologies, content and human rights

Sameer starts off by framing the discussion, about the intersection of human rights, videography, and photography. Makes reference to last-night’s breakup of the occupy Oakland movement, with visual content, almost in realtime, and the sort of content that we would normally never see. Asks the panelists to pinpoint the moment they saw the shift of video and human rights to a big, participatory issues.

Hans: For us, we saw right ahead of the elections in Egypt last year (we had no idea what was going on) a huge number of signups. And there were 10,000 videos that came out, and we saw that people were trying to document what was happening, because there were no independent observers. And it hit us in the stomach, it was a very emotional moment. And just before we came up here, I was watching videos from Syria, where people had set up their own wifi, singing, chanting.. it’s hard to explain until you see it yourself.

Steve: The moment for me was when there was a screenshot from Burma on the front page of the Wall Street Journal. But of course, the actual watershed moment was the Neda video, which then won a Polk award, and you couldn’t figure out who was the originator, because it had been uploaded and taken down, and uploaed again and again. And from the Arab Spring, some of the videos were so violent, and so graphic, and this really brought it to another level, where it wasn’t just a video, but it was evidence. It was evidence of wrongdoing.

Thor: It was, for me, the fall of the Berlin Wall. And there was no wall falling, it was being torn down. And there was no social media, but it inspired all these people to say, they did it, we can do it too–and how powerful an image can be. I remember when I had the first Razr, when people were sending around these images of human rights violations in Latin America.

Sam: I was going to mention Burma, again. And people would ask, where is that clip of NHK footage, the only clip of what had happened in 1988 suppression. And now, following the Saffron Revolution, that scarcity problem as gone away. And you have this huge expansion of people who are involved in documenting these human rights issues. The people who run the Syrian Revolution Facebook page, who are anonymous, might not even be based in Syria. The person who collected 3,000 videos from Iran and gave them context, was based in New Jersey. And I think about the one or two videos from Sudan, and from the CAR, whose voices are lost, because their message doesn’t surface.

Sameer: Let’s pick up on that issue of scale. The quote of Victoria, 48 hours a minute, that will only grow. That’s a vast proportion of how we communicate, and interact, and what we’re seeing. You’ve identified these great challenges–the scale and realtime nature.

Thor: There’s a difference in categorizing and cataloguing video for the record, and for video as advocacy, to make governments act or to create advocacy. And we should be clear about these differences. Witness is going to something different tan hose who have an agenda at the UN level. So let’s move this away from a nebulous discussion.

Sam: One of the big issues we face when looking at the range of users, there’s an intersection between those creating the spaces, and what role can these platforms play in helping users create these videos safely, use them as evidence, or disseminate them widely. There’s a conversation about how users create footage with intentionality, or those who create it and we impart human rights value upon them.

Steve: When we look at the differentiating value on YouTube, when someone take a look at it, our community reviews it. There’s an EDS exception–education, documentary, or scientific–then we’ll make an exception and leave it up. If it’s incredibly gruesome and difficult to watch, we’ll create an age gate. And if it’s really, really gruesome, we’ll create an intermediate page as a warning. From a company perspective, we have a huge discovery problem, which is how do you find this content when it’s on the site, and we need to balance between recency and relevancy. Recency is important for news, but relevancy is something you discover later in the news lifecycle. And every day we run news content from the site–we work with  company called Storyful to find that breaking news. And we don’t conflate the two, news and human rights, but there is often an overlap. And so putting it into context is important. The future of content on YouTube and other platforms is really a question of how can curation happen, and curation is a matter of preference. So, what is the best or most relevant comes down to individual voice. So we do a good job incentivizing viewing and uploading, but not yet having tools to incentivize curating. So that’s something we’re looking at for next year. How can we incentivize curation, and create a bunch of different tools and lenses about what content should look like, and how to share that more easily.

Hans: We’re 12 people, so we have a slighty different perspective. We don’t sell ourselves as a destination site, because we believe in free distribution of content, so we want people to share wherever their home is on the internet, rather than driving traffic to our site. But when the election happened in Egypt next year, we started communicating with bloggers and activists, asking if we could help.And their main request was help us get visibility, help us distribute this. So we set up a page for all the videos coming in from North Africa and the Middle East. But with regards to safety, we could only register those coming in from geotags from Egypt, or people who had registered as such, but we didn’t have the resources to curate for those who hadn’t given us that baseline data. We’re doing our best, but it’s a tricky question for a small company to solve.

Sam: A lot of video shot on mobile leaks a lot of information about where you are. And there are Egyptian activists who chose to show their locations–Wael Abbas (@waelabbas) wanted people to see his location, so he could demonstrate where there were roadblocks. You know, if you’re in Syria, you don’t want to be geolocated, you might not want to be visible in the frame, and you might not want to have the background imagery in the frame. And with certain things, like anonymizing your IP address, is easy, but anonymizing video is a big challenge. You have to go into your FinalCut Pro or whatever. This is a major challenge. As Alaa (@alaa) said, activists are only 1% of the users, but these issues should be relevant for the other 99%. For those not in Syria, those who might not be out, those who are subject to domestic violence, for example.

Hans: We don’t force people to give away information when registering and broadcasting. But the challenge is, how do you educate the user? Ramy Raoof (@ramyraoof) did tremendous work creating manuals in Arabic about how to be safe using Bambuser, and distributing this in many ways. We want you to be safe using Bambuser. I don’t know if it is relevant, but we’re a Swedish company, and we have fair legislation, we’re a democracy, and our servers are protected under Swedish law. One request we’ve had is that, they have poor connections, or no connections, but we want to use Bambuser. We do live broadcasting, but now, starting next week, we’ve created an offline opportunity for recording and uploading. And really, this is 99% percent in response to our users in North Africa and the Middle East. We want people to see Bambuser as a full video service, whether you have a data connection or not.

Thor: Witness put cameras everywhere, but of course, there are so many places they cannot put cameras, they can’t fund being everywhere. A great example of places to have cameras would be Cuba, North Korea. And this video takes a very long time to travel out, and it become difficult to know who took it, and what you’re really looking at. So one suggestion for companies might be to start assigning by topic–can, rather than curating a whole bunch of news, just have a place for raw footage from a certain location, like elections. So rather than look at things like armed conflict, which is difficult to look at, video, and a 5 minute clip from one company, can open a Pandora’s Box about what really happened.

Steve: On YouTube, you can be whatever you want. You can be Jellybean202. So then the question is, did this happen, and is it relevant. And we have this challenge of verification of when and where did this happen. And as YouTube, we take the stance of, we’re a platform. The next layer of context should come from organizations with a better ability to understand context, like Witness, like the media. We launched a new site, called YouTube Trends, that is a water cooler of what is happening now. We look at what is a rising video on YouTube, and try to package it with some context, to help people understand.

Sameer: Coming back to visual privacy–your written identity for example, can be obscured to some degree, though there are linkages beginning to happen–now your YouTube account needs to be linked to a Gmail account.

Sam: The human rights case illuminates some of the other issues. We’re the 1%, the highly unusual requests, at the extremes of use. So, we’ve been involved with the Guardian Project, to create an Obscura Cam, to anonymize the image in the frame, and the underlying data. It may not matter if the uploader is safe, it matters whose image is seen. It was designed for human rights activists, but what we saw when we did the alpha release was that people wanted to use it to take a picture of their kid on the playground, and obscure the pictures of the other kids, because parents hate having images of their children up on the internet. And we know that most people will generally choose not to obscure their images, but some people will. And the other tool we’re developing is InformaCam–which allows people to capture as much data as possible, to use as evidence. I’m here in Syria, I’m in Homs, it’s a Friday. We want to give users this control.

Steve: There’s this question, too, about who has the burden of responsibility about identity. We give users the ability to write in, to say they’re in a video, don’t want to be, and have it taken down. But sometimes, you don’t know you’re on a video until it has 300,000 views. Am I supposed to check YouTube everyday to see if I’m on YouTube?

Sam: I think this is an acquisition challenge. I was looking at the YouTube editor, it would be easy to give users the opportunity to pixelize as they upload, for example.

Thor: People know what they do might end up on camera, and they’ll increasingly know this. It was clear after the Arab Spring that these tools had not been designed to help activists on the ground. Activists ended up using them. There are things that can be added to these platforms to make them a lot faster, to action the issues–but ultimately, much like what happened when Wikileaks released documents that exposed people who were providing data about human rights. People on the ground need to know this sort of thing might happen, and the education part is so important. And knowing to look at a space on the website, that’s great for those with a 12th grade American education, but if all you know is how to press record and upload, that’s not helpful.

Sameer: The other outstanding issue is facial recognition. This is a tool that is now being integrated into all these tools, like iPhone. The new version of Android can look at your face and unlock your phone. These things are not quite domesticated, but they’re coming in. Yet they’re not part of the media literacy that we have yet. So there’s a broader cultural question.

Sam: There’s a really interesting facial recognition challenge from Carnegie Mellon University. They just compared people on Match.com to Facebook and found it was quite easy to identify people. And they identified 1/3 of people on the street to Facebook. And then, they reviewed the data aggregated in people’s social media profiles, and were able to identify their Social Security Numbers. So I’d put it out as a challenge to those in the room as a means of thinking about this.

Thor: People who go on the streets, and are willing to die, often say, by all means, go put my image out there, I’m willing to to this. And if the Bahraini government goes ahead and persecutes them, it gives us yet another opportunity to highlight that hypocrisy and violence. I’m a believer of the more, the better–when you start using the argument of privacy, that’s the same opportunity that governments user to censor. more information may be messy, but censorship is invariably messier.

Hans: Of all the hundreds of thousands of videos from the Arab Spring on Bambuser, we yet to have one takedown request. It’s possible many people don’t know they’re on video. But this is where the importance of education comes in. The cameras are all around, they’re closer than you think.

Sam: People choose to take risks, but at the moment, they don’t really have options about how to take them. It’s important to remember that the Arab Spring is really jsut the tip of the iceberg on human rights. Sex workers fighting police violence in Macedonia, elderly who face physical abuse, sexual violence in South Africa–these are the other stories. I think people need additional controls of how these images are used. I do agree, in the broader sense, that more cameras are good. And that informed consent and privacy are inversely proportional to power.

Steve: If it’s a public figure in a public space, versus a private figure in a private space, that influences how we deal with content. WE’re not in the business of determining veractiy. We are only in the business of trying to determine the intent of the uploader from the contextual clues about the uploader–i.e. the glorification fo bullying, versus the documentation of bullying.

Thor: Again, the more the better on the question of veracity. It’s no longer just the uploader and the producer. It’s going to be exposed to thousands and thousands of viewers, who can peel back those layers. Two things that come to mind: April 11th, 2002, when Chavez was removed from power for 48 hours. There are still 19 unsolved murders, and the video evidence seems to show the government murdered 19 people and got away with it. After the fact, someone with $4,000 was able to show how a documentary video had been completely manipulated. In Guatemala, there was the instance of the attorney who made a video accusing the president of his murder, but the reality was that it was an arranged suicide, and through sleuthing, this was found out, and a suicidal coup was averted.

Sameer: We’re talking about what the companies should do, but this is a big ecosystem. Given what we’ve talked about, including user education, there is an emerging sense of responsibility that sits with companies. So what should we see?

Sam: Start by doing human rights assessments on risks of new technologies. Start out by testing out the cases to see what happens. Be in touch with the human rights communities to see what happens and think about education. And look at innovation happening in the human rights technology community, to think about how that might work in your own technology.

Thor: There are thousands of human rights organizations. Diversify as much as possible and get as many voices to the table as you can. Don’t just go with one or two–there are politics in the human rights community, just as anywhere else. Trust, but verify, and diversify to do so. Make these platforms more accessible to people locally. Talk to people on the ground about how to use these tools, you might create great tools, but barriers to access may be too high.

Steve: We do try to do these things, so the biggest thing for us is just listening. Trying to hear about what the priorities of the human rights community are. So, for example, with the challenge of offering pixelation, we want to do this, but it creates a business flow challenge by slowing uploads.

Hans: Listening and communicating is key, and we work with our users to hear about what their needs are, and we try to reply to every email that comes in. We can’t figure it out ourselves. The people who are actually doing it know what their own needs are.

Sameer: There are different departments in companies; product, PR, and the like. As companies grow and scale, it should be important to bring people from across the spectrum. This should’t be CSR, this should be embedded across the company, from engineers to everyone else.