Subscribe

FairByDesign · The Reconstruction Test

Could you reconstruct what your AI did?

Not explain what the model is supposed to do. Not show your AI policy or your risk register. Pick one real, consequential decision an AI system made or shaped in your organization, and answer six questions about it honestly.

Two minutes. Nothing you answer is sent anywhere or stored. The six questions come from FairByDesign’s field paper Can You Reconstruct What Your AI Did?

1. What happened?

The event itself: outputs, logs, transactions, actions taken.

2. Why did it happen?

The rationale: inputs, context, prompts, the policies in force, the decision record.

3. What influenced the outcome?

The causal chain: data sources, external tools, upstream models, human inputs.

4. What controls existed?

The governance state: guardrails, approval gates, access limits and policy controls actually active at the time.

5. Could intervention have occurred?

The stop authority: escalation paths, override rights, human-review records.

6. Who is accountable?

The ownership: assignments, accountability maps, authorization records.

A self-check, not legal advice, an audit or a certification. Reconstruction is not a mandate for indiscriminate logging: evidence should be proportionate to the significance of the decision, kept only as long as justified, and protected against reuse.