Key themes
- Information overload: Journalists now receive massive volumes of user-generated content and must filter authenticity rather than hunt for stories
- Social media as a newswire: Twitter and real-time platforms have become primary sources, requiring journalists to assess credibility of sources on the fly
- Open-source verification: Free tools (Google Maps, Wolfram Alpha, Spokeo, reverse image search) enable fact-checkers to cross-reference claims, locate sources, and debunk fakes without special access
- Visual geolocation: Analyzing distinctive features in videos and photos (railings, curbing, shadows, landmarks) can pinpoint locations and verify whether content matches claimed contexts
- Human judgment: Algorithms encode binary rules, but truth is contextual, emotional, and fluid—it requires human interpretation even as technology accelerates fact-checking
Case studies
Hurricane Sandy fakes (2012)
When Hurricane Sandy hit New York, journalists had to sort through unprecedented volumes of user-generated media, including old photos reposted as current, composite images merging previous storms, and clips from disaster films. Nolan describes verifying an Instagram photo of flooded Avenue C by identifying the posters (New York food bloggers) as credible sources, then cross-referencing visual details. The key move: trace the source.
Rita Krill video verification
A video showing someone reacting to a backyard explosion during a storm looked authentic but required verification. Using only free internet tools and a YouTube username, the team:
- Located potential Rita Krills in multiple states using Spokeo
- Checked weather records (Wolfram Alpha) for each location on the upload date
- Narrowed to Florida (thunderstorms that day)
- Searched white pages and found addresses
- Used Google Maps to locate a house with a pool matching the video background
- Cross-referenced physical details (umbrella, white lilo, pool edges, trees) from the video against satellite imagery
- Confirmed the upload as genuine
Result: 20 minutes using free tools, sitting in an office in Dublin.
Syrian conflict video (bridge identification)
Journalists investigating alleged war crimes in Hama, Syria needed to verify a video showing bodies being dumped from a bridge. Three local sources gave conflicting information about the bridge's existence. By analyzing the video for clues (railing design, curbing pattern, blood in river indicating flow direction, shadow direction indicating bridge orientation, and visible divot on riverbank), the team systematically checked every bridge in the region using Google Maps.
Method: East-west orientation, black-and-white curbs, specific railings. Found the bridge by switching to satellite view, zoomed into Google Maps photos linked to the location, and cross-referenced architectural details. Conclusion: The bridge existed; the source providing geographic/hydrological details was most credible.
Notes
This 2012 talk predates modern deep fakes and AI-generated content, yet the verification principles—source tracing, cross-referencing, metadata analysis, geolocation—remain foundational. Nolan's emphasis on human judgment ("truth is human") is particularly relevant as automated systems proliferate. The talk exemplifies how journalists adapted verification workflows to the social-media era, building on established investigation methods but at internet speed.
Storyful, founded by Nolan, became a major player in open-source verification for news organizations and later Agence France-Presse.