Start here
Your first five days: inventory, risk triage, named owners, and a 30-day plan. Start here if you have no idea where to begin.
Ten questions that surface your single biggest governance gap before you scale an AI system.
Every key governance term defined in plain language. The reference you bookmark and cite.
Risk
Thirty named AI risks across six categories: the reference map the whole channel hangs on.
A 12-question, five-minute triage that ranks which risk categories expose you most.
A fast check for autonomous and frontier systems: the risks that act without asking.
Why the most advanced AI systems are the least governed, and what to do about it.
Compliance
The verified world map of AI regulation, from the EU AI Act to Brazil, Africa, Asia-Pacific, and the Middle East, with a first action for every regime.
The EU AI Act, ISO/IEC 42001, and the NIST AI RMF on one sheet: law, proof, and method.
What to tell users when they interact with AI, anchored to the EU AI Act's transparency obligations.
Twenty checks that tell you whether your governance would survive an independent audit today.
First-party, second-party, third-party: which level your board or regulator actually requires.
Operations
What to do in the first sixty minutes when an AI system fails in production. Contain before you explain.
The front door of governance: capture every AI system before it ships.
An eight-question gate for the AI you buy, not build: the largest exposure for most organizations.
When a human must stay in control, and how, per the EU AI Act's human-oversight requirement.
Who is accountable for what: the org-design layer most frameworks skip.
The cadence that keeps governance alive after launch.
Ethics and people
Twenty questions across five sections that score how defensible your AI decisions are.
The seven-principle value spine, including the one no one else names: truthfulness over agreeableness.
How ready your staff are for the AI they already use. The EU AI Act now expects literacy.
The ladder from unaware to fluent, and where your organization sits.
The rights-based check public-sector and global organizations require.
Why human rights is the lens that makes AI governance defensible everywhere.
Flagship reading
The quarterly flagship publication: the five forces reshaping AI governance, from principle to practice.
For engineers
The same instruments, as machine-readable schemas and a runnable test suite. Importable, forkable, citable by stable id. The frameworks of AI GOVERNANCE CHAIN™, shipped as working data.
All 30 named risks and six categories as structured data. Load it into your risk register, tag your AI inventory, or cite risks by stable id.
Ten paired-prompt tests covering all eight sycophancy patterns, with a working Python scorer. Run it against your model today.
The 16-question, 64-point instrument as a data schema. Embed the score in your own dashboards and pipelines.
One append-only entry per AI-influenced decision, with the fields an auditor accepts. Drop it into your repo.
One file per AI system, with the risk-tier routing logic built into the structure.
The developer index: what each file is, a copy-paste quick start, the canonical-ids rule, and the free-to-paid boundary.