After a quick break I’ve made my way to Copenhagen for the iCS Symposium “Locked Out of Social Platforms”, and the first panel of the day starts with a paper by Nicholas Proferes. His focus is on how power is manifested in the platform affordances of social media: these include affordances such as the persistence, visibility, spreadability, and searchability of content.
Nick focusses here on the case of WikiLeaks’ release of the Podesta Emails, from a hack of Hillary Clinton’s campaign chairman John Podesta’s email. These were released over thirty batches, starting just after Donald Trump’s infamous Access Hollywood recording broke. Each batch was associated with a new hashtag (‘#PodestaEmails’, ‘#PodestaEmails2’, etc.) in order to generate a new trending topic for each batch; had they all used the same hashtag only the first batch would likely have trended.
The project captured every tweet authored by WikiLeaks’ account, using a clever Web scraping approach for historical tweets that Nick and his colleague Ed Summers have made available on GitHub; it also accessed historical trending topics information via Trendogate.
WikiLeaks had never serialised releases and associated hashtags in this way before; its approach maximises spreadability, visibility, and persistence in the weeks before the 2016 U.S. presidential election, as the Twitter data show. Some of the trending of these emails was facilitated by the Russian Internet Research Agency bot farm, and suspiciously the hashtags usually began trending in Vietnam first.
But persistence, visibility, and spreadability were also achieved beyond Twitter as mainstream news organisations picked up on these releases (and on their trending); these stories were then in turn retweeted by Donald Trump and other accounts connected to his campaign. We may understand this as an indication of the symbolic power of this story, but that power also depended on WikiLeaks’ gaming of Twitter’s trending topics system; this is building on informational power in Sandra Braman’s definition, then.
