Crowdsourced Images in the Boston Marathon Attack

The next speaker at ECREA 2014 is Anssi Männistö, who shifts our focus to the Boston Marathon bomb attack. Mobile social media played an important role in covering this attact: tweets and mobile media were no longer just sources of information, but also tools to very facts and photos and to identify potential suspects, through image recognition software and other facilities.

In Boston, journalists rapidly discovered the first reports and images of the attack from Twitter, and soon came to use them in their own coverage. Such material was then used in official investigations, unofficial hunts for the culprits, and in the media coverage. These each drew on a massive amount of mobile photos; on the real-time publishing of such content in social media; and on crowdsourcing of activities through social media.

Law enforcement used some 10 TB of images and videos provided by the public; they actively encouraged the public to submit their material. Unofficial investigations by the public also began, and self-styled investigators shared images of potential suspects, many of whom turned out to be innocent. Some such images also made it into the mainstream media. The subsequent manhunt for the culprits utilised infrared camera footage and embargoed these images from use in social media until after the capture of the suspects.

Subsequent studies of the event documented the crowdsourced image verification practices some of this work engaged in. This drew on volunteers pooling their information and using background data such as Google Street View to verify footage, trawling social media profiles to verify the identity of people posting content, and engaging in quasi-journalistic activities to confirm the history of content items.

Mainstream media and marathon organisers in Boston used their social media accounts extensively to cover the bombing and its aftermath. The Boston Globe covered the bombing more intensively for longer, while the Boston Marathon organisation switched back more quickly to business as usual.

What we see emerging here is a new network of surveillance enabled by the ubiquity of mobile cameras. We are also seeking the willingness of people to engage in the crowdsourcing of event coverage and investigation.