Twenty-nine modules, one assessment.
Each module answers one question about an AI system, from why fairness matters to whether a release may ship. They follow four phases: Diagnosis, Decision, Intervention and Enforcement. Turn the wheel to explore them, or open the one your project needs.
Arrow keys move along a ring, up and down move between rings, Enter opens a module, Home turns the wheel back to the start.
How to read the wheel
Four phases, in the order you work.
The inner ring holds the 4 phases, from START clockwise. The middle ring holds the 17 steps, the rim the 29 modules, each as tall as it is important. Point at any part, or turn the wheel by its edge.
- Diagnosis
- Decision
- Intervention
- Enforcement
Three ways into the modules
PulseLive
A fairness triage reading in one sitting: declare what you bring, get a plain-language opinion and a report, without writing code. A first read, not an audit.
On the Platform pageFairness NavigatorBeta
The guided path: the steps in order, with gates and an explicit decision at each stage. Several modules open from its steps and hand their results back.
On the Platform pageModules on their ownLive
Run any module à la carte, in the order your project needs. Work is saved with the assessment.
On the Platform pageAll 29 modules, phase by phase
Diagnosis
The Diagnosis phase establishes whether AI fairness is materially relevant for a given system and, if so, where the primary risks are likely to arise. It frames the assessment context by defining scope, constraints, and plausible harm mechanisms, ensuring that subsequent decisions are justified, proportionate, and evidence-based.
Mandate & Scope
Business & Policy
Context & Risk
Decision
The Decision phase defines what fairness means in the specific system context and for whom it must be guaranteed. It locks normative choices, population scope, and measurement criteria, preventing arbitrary metric selection or post-hoc justification later in the lifecycle.
Populations & Intersections
Bias Identification & Prioritization
Intervention
The Intervention phase determines how the system may be legitimately changed to address identified fairness issues. It relies on causal and counterfactual reasoning to ensure that interventions target root causes of unfairness without undermining legitimate system behavior.
Counterfactual Test
Intervention Plan
Techniques
Enforcement
The Enforcement phase ensures that fairness requirements are validated, enforced before deployment, and sustained over time. It converts analytical findings into binding go/no-go decisions and embeds fairness into organizational governance and operating practices.