Cognitive War 31: The Enemy in the Briefing Pack

We move now to the emergence of indirect prompt injection as a sophisticated tool for cognitive warfare, where malicious instructions are hidden within data to subvert AI-assisted decision-making.

These attacks function through instruction laundering, a process where a hostile command is “washed” into the AI’s professional tone, making the resulting bias or manipulation invisible to human reviewers.

We argue that this creates a dangerous verification gap, as experts often lack the time or raw evidence needed to detect when an assistant has been covertly hijacked. To counter this, we propose a governance framework centred on strict permission boundaries, deterministic safeguards, and the preservation of meaningful human oversight.

System designers must ensure AI assistants remain tools for information rather than unauthorized decision-makers that strip humans of their agency and accountability.

Full article available here

Cognitive War 31: The Enemy in the Briefing Pack

·

Aug 18

Imagine a Monday morning at 8:54. A hospital procurement team receives the final safety file for a new clinical logistics platform. The document is eighty-three pages long. Their copilot reads it first.