Smashing Security The Epstein Files didn’t hide this hacker very well
9 snips
Feb 5, 2026 Tricia Howard, cybersecurity marketing leader and former security researcher, joins to unpack sloppy redactions that let AI and public data reidentify an alleged hacker. They also dig into a senior official's accidental ChatGPT upload and the rising insider risk as AI tools make leaks easier. Light picks of the week add a fun cultural break.
AI Snips
Chapters
Books
Transcript
Episode notes
Gmail Replica Made The Files Searchable
- Graham Cluley describes jmail.world which mirrors Epstein's Gmail so users can search millions of pages easily.
- He recounts how quickly the community identified the redacted individual using simple biographical clues and AI.
Reputational Damage Outlasts Facts
- Allegations in leaked files can damage trust even if unproven, especially when tied to major cybersecurity vendors.
- Graham Cluley notes trust is hard to build and easy to fracture, leaving clients nervous and boards asking questions.
Treat Biographical Bits As Identifiers
- Organisations must rethink redaction processes for the AI era and remove identifying breadcrumbs, not just names.
- Graham Cluley advises treating collections of biographical facts as identity information when releasing documents.
