David Haines family photo. Written by Sam Gregory, Program Director at WITNESS. This post was originally published on the WITNESS blog. In the last month, many people in the US and worldwide have been exposed to videos showing grotesque acts of violence, notably the killings of two journalists, James Foley and Steven Sotloff, and British aid worker David Haines by ISIS. Although sadly not exceptional in their graphic violence for those who watch the Twitter-stream of human rights and crisis footage, these received significant prominence in the mainstream media and have ignited an important conversation about roles and responsibilities in relation to graphic imagery and gross violations of human rights. In response to the circulation of the videos created by ISIS, social media platforms responded by taking down the footage, and individual citizens, including family members of those murdered, started hashtag-driven movements and personal appeals around their own decision not to watch and encouraging others not to watch. [Featured image from Twitter user Sajad Riyad – part of an effort by supporters and colleagues of British aid worker David Haines to circulate imagery of Mr. Haines “as he would want us to remember him” rather than the video of ISIS murdering him which became an international news story, September 2014.] Within the community of people WITNESS works with, these news events point to critical questions for commercial social media platforms, and that we need to engage with as citizens, users of social media, and consumers of information and imagery. Although often we perceive of spaces like YouTube and Facebook as free and open public spaces, they are not. In Ethan Zuckerman's analogy they are more akin to malls, largely driven by the commercial imperatives underlying them as much as by a public space role. They are governed by relatively broad guidelines on what content is acceptable or not (that do allow a fair amount of discretion to the platforms), and differing perceptions from viewers and internally on what their audience and purpose is. While it is not ideal that decisions on free speech are made by commercial platforms, given the role these platforms play in circulating critical information, WITNESS advocates for controversial content to be shareable and to remain on these platforms, and for as much clarity and consistency as possible on when/why material is removed. YouTube, for instance, generally allows graphic footage only in contexts where it is framed as evidence of rights violations, rather than a glorification or validation of the act. So, in the case of the James Foley footage, YouTube referred a reporter for Slate to their content policies in relation to gratuitous violence, incitement to violence and hate speech: “YouTube has clear policies that prohibit content like gratuitous violence, hate speech, and incitement to commit violent acts, and we remove videos violating these policies when flagged by our users,” a YouTube spokesperson told me over email. “We also terminate any account registered by a member of a designated Foreign Terrorist Organization and used in an official capacity to further its interests.” Holding to these types of approaches consistently, irrespective of who is involved in violations, will help establish a more transparent standard. Even with clear parameters, it is the case that some footage on YouTube and on other social media spaces that has important value for justice and accountability will get taken down – either because it is not framed as evidence, or because of human judgement calls on whether material fits within the company's guidelines. Footage that shows human rights violations is often graphic and uncomfortable, sometimes re-victimizing and is frequently subject to being arbitrarily or correctly removed either by administrators or concerted take-down attacks by outside actors (a relatively common problem on many social media spaces). When we analyzed the playlists of citizen video that have been shared on the Human Rights Channel (which is hosted on YouTube), we found that of the almost 6,000 videos showing rights violations that we have shared, almost 5% are now missing. This could mean they were deleted, removed or made private. One idea that has been circulating is that of a digital “evidence locker.” The “evidence locker” would make sure that powerful but offensive citizen media related to human rights is downloaded and saved. This would be done in a way that preserves metadata and other important video information, so that it can potentially be used in future prosecutions and investigations by NGOs and human rights actors even if it is rapidly deleted on a social platform.