TCI Insights

Threat Analysis

Threat landscape assessments, adversary techniques, and risk analysis.

All Insights
Threat IntelligenceLLM Obfuscation & Redaction8 min

When LLMs Leak: Obfuscation Failures in AI-Assisted Redaction

LLM-powered redaction tools can fail in subtle, dangerous ways — reconstructing supposedly removed PII from context, leaking information through embeddings, or hallucinating redaction where none occurred.

David Greenhill
David Greenhill
Read more
Threat IntelligenceLLM Obfuscation & Redaction7 min

Prompt Injection as a Redaction Bypass

Documents processed through LLM redaction pipelines can contain adversarial text that manipulates the model into preserving sensitive information. This attack vector is unique to AI-powered redaction.

Michael Peterson
Michael Peterson
Read more