If a decision affects you, you can question it.
An AI system that cannot be argued with is a liability, not an efficiency. Open to Challenge guarantees a clear, visible path for anyone affected by an AI-influenced decision to ask for a human to look at it again.
What contestability actually looks like
A visible path, not fine print
Anyone affected by an AI-influenced decision has a clear, simple way to ask for it to be reviewed by a person.
A guaranteed human look
Challenges get a genuine human review, not just a form that disappears into a queue.
Logged and reviewed
Challenges are logged, so patterns of complaint actually change how a system behaves over time.
Take it seriously by default
The default response to a challenge is to look at it properly, not to defend the system first.
Contestability is what separates automation from arbitrariness
A decision that cannot be questioned, right or wrong, feels arbitrary to the person on the other end of it. This guardrail is what keeps automated decisions feeling fair, and gives you an early warning system for when a decision-making process actually needs fixing.
Grounded in
- Australia’s Voluntary AI Safety Standard, guardrail 7 (contestability)
The rest of the framework
All ten guardrails run through every engagement. Here are the other nine.
Frequently asked questions
Isn’t this just a complaints process?
It is a subset of one, specifically for AI-influenced decisions, with a guaranteed human review, not just a form that goes into a general queue.
What kinds of decisions actually need this?
Anything where being wrong has a real cost to the person on the other end: a declined application, a de-prioritised request, a pricing decision.
Find out where your governance actually stands. Governance is one of six things the free audit scores.
Get the free audit