Book excerpt looks at NSA's Big Awesome Graph initiative, an attempt to map people's social networks out of an enormous trove of intercepted communications
Barton Gellman / Wired :
Context & Ripple Effects
Barton Gellman — one of the journalists who received the original Snowden trove — has been serializing his book Dark Mirror, with an adapted excerpt already published two weeks before this piece. The new Wired excerpt names a previously less-visible program, Big Awesome Graph, that sits on top of the collection systems already documented in this coverage arc.
That arc runs from mechanics to analysis: The Intercept explained how XKEYSCORE collects and searches intercepted Internet data, and a later Snowden leak showed GCHQ's data-mining techniques doing similar work at the UK agency. Big Awesome Graph completes the picture by describing what the NSA builds once communications are collected — a map of who knows whom.
First-order effects
- Gellman's book puts the NSA's social-network mapping effort into the public record under a proper name, shifting it from an inference drawn from document fragments to a formally described program attributed in print.
- Readers of the Dark Mirror excerpt can now connect the NSA's collection tooling directly to its analytical output — raw intercepts becoming relationship graphs of identified people.
Second-order effects
- The GCHQ material in this coverage shows allied agencies running comparable data-mining pipelines, so disclosure of the NSA's graph initiative invites scrutiny of whether partner agencies maintain parallel social-mapping programs.
- Each republication of Snowden-derived material through books rather than news cycles extends the shelf life of these disclosures, keeping surveillance programs in public debate long after their initial leak.
Third-order effects
- The pattern across XKEYSCORE, GCHQ's mining techniques, and now Big Awesome Graph points to social-network analysis becoming the default analytical frame for bulk-intercepted communications — intelligence value measured in mapped relationships rather than message content.
- If agencies continue deriving relationship graphs from mass collection, the unresolved policy question this coverage keeps circling is how domestic oversight regimes govern metadata-derived profiles of people never individually targeted.
The trend: Signals intelligence is consolidating around mapping social networks at scale, with bulk-collection systems feeding graph programs that turn intercepted communications into maps of human relationships.