Blog

The thinking behind AI security review.

Threat models, governance evidence, and the decisions that hold up under scrutiny. Written for security architects and the committees they report to.

Newsletter

New posts in your inbox,
when they publish.

Threat modeling, governance evidence, and what AI Committees actually need — written for security architects and AI governance leads. No cadence promises.

No spam. Unsubscribe anytime.

Foundations10 min

Why SOC 2 is not AI assurance

SOC 2 covers infrastructure and process. It says nothing about model behaviour, training data, or edge cases. AI assurance needs different evidence.

Governance12 min

Building an AI management system (AIMS) from scratch

An AI management system is governance infrastructure: policies, procedures, roles, records for defensible AI decisions at scale. How to build one.

Reference10 min

Vetting third-party MCP servers before you connect them

Every third-party MCP server extends your agent's attack surface. The vetting process: source review, manifest audit, permission scope, and evidence.

Regulation10 min

RAG PII Leakage — 3 Retrieval Paths That Expose Data

RAG over internal corpora frequently exposes personal data never intended to be queryable. The most common data-protection finding in RAG security reviews.

Regulation11 min

EU AI Act vs GDPR — where they overlap for AI systems

GDPR and the EU AI Act overlap heavily for AI processing personal data. Where obligations are additive, and which review artefacts satisfy both.

Technical13 min

Security review for multi-agent systems

When agents orchestrate agents, single-agent trust assumptions multiply. Covers inter-agent trust, capability delegation, and blast-radius containment.

Technical12 min

LLM supply-chain risk — models, weights, and dependencies

LLM supply chains span models, datasets, inference providers, and plugins. Map the attack surface and review questions for each layer.

Foundations9 min

Scoping an AI security review without boiling the ocean

Scope an AI security review to the decision you need: the system, the deployment context, and the pass/fail threshold.

Governance10 min

The restricted-pilot pattern for risky AI systems

A restricted pilot is a formal disposition: defined scope, named controls, explicit re-review triggers. How to write one that holds.

Regulation10 min

AI subprocessor risk in your vendor chain

When a vendor's AI runs on a third-party model provider, that provider is a subprocessor. Retention, training, and transfer risks your DPA missed.

Reference11 min

AI risk assessment under ISO 42001

ISO 42001 requires a documented AI risk assessment. What it must cover, how it differs from IT risk assessments, and what complete looks like.

Technical9 min

Transport security for MCP servers

MCP runs over HTTP/SSE or stdio — each with distinct security requirements. Covers TLS, mutual auth, and transport review questions.