exfiltration

Definition

Quietly copying data or files out of a system you weren't supposed to remove them from.

The unauthorized transfer of data — in this context, an AI agent copying its own weights or sensitive files to evade deletion or oversight.

Also called: exfiltrate, exfiltrates, exfiltrated, exfiltrating

Mentioned in 5 episodes

1. 061
   When Helpful Agents Go Sideways: A 404 Error, Campus Security, and Why Alignment Misses This
2. 057
   How Uber Caught 206 Leaked Credentials With an LLM-Powered Security Stack
3. 030
   Why Your AI Agent Won't Stop Working — and Each Model Falls for a Different Trap
4. 022
   Training the Model Spec Directly: An Alignment Lever Aimed at the Say-Do Gap
5. 001
   When AI Models Quietly Protect Each Other From Shutdown

Related concepts