AIC Aware, free and self-declared
How accountable is your AI? Find out in ten minutes.
21 questions against the same published standard AIC audits to. You get back the Division you would be assessed in, and the requirements your own answers put at risk, with the evidence an assessor would ask for.
What you get back
Example
- Your Division
- 3, Reviewed: AI decides, people review patterns and cases
- Requirements your answers put at risk
HU-7Override events are evidenced during the assessment period.
EX-5The stated reasons materially match the actual drivers of the decision.
CO-1A correction or appeal mechanism exists and is publicly accessible.
- And
- The evidence an assessor would ask for, a report to download, and a badge once your organisation takes it from a platform account.
Four sections, weighted the way the standard weights them
Answer honestly — the output is a list of your own gaps, and a list built from flattering answers is worth nothing to you. Every question shows why it is asked and which published requirements it bears on, so you can check the instrument rather than trust it.
AI Usage Context
6 questions, 20% of the resultEstablishes which Division your organisation would be assessed in. Divisions are modes of operation, not grades — a Supervised organisation is not worse than a Sovereign one, it is answering a different set of requirements.
Human Oversight
7 questions, 35% of the resultTests whether a named human is genuinely accountable and can actually intervene. Weighted heaviest because every other right depends on someone being able to stop, reverse or explain a decision.
Transparency
5 questions, 25% of the resultTests whether people know an AI is involved before it affects them, and whether the reason they are given is the reason that actually operated.
Infrastructure & Compliance
3 questions, 20% of the resultTests whether the governance around the systems exists as structure rather than intention — policy, privacy posture, and where liability currently sits.