Public casebook
DAI Casebook — Real investigations, documented under the DAI Methodology.
Public entries from TRST investigations, written to show what the evidence supports, what it does not support, and what remains unknown.
The record includes unresolved questions, corrected failures, and the evidence that changed the methodology.
One Writable Word Produced a Confident False Accusation
A genuine crash photograph became "AI-generated media" after an attacker wrote the name Midjourney into an unsigned metadata field. The pixels did not make the accusation. TRST did — and that failure changed the evidence rules.
The Photo Meant to End the Rumors — and What Verification Could Actually Say
A photograph released to show a U.S. senator was alive spread to twenty newsrooms overnight. That breadth looked like corroboration. It wasn't — and separating the two is the whole discipline.
The Image Was Obviously Fake. The Claim Was Unfalsifiable.
A viral photo of "vampires at the White House" fooled almost no one. The story attached to it — posted, then instantly deleted — couldn't be checked by anyone.
One Photograph, Four False Lives.
An authentic 2009 protest photo has spent seventeen years being recycled as evidence for claims it never supported — including opposite ones.
The Image Was Real. The Claim Was Not.
A viral World Cup image showed a real broadcast moment — but the claim-bearing face region was altered.