Adversarial Reference Cases

Status: Working draft

Adversarial Reference Cases are designed to test where a framework fails or whether it fails safely.

They are different from ordinary Reference Cases.

Reference Cases ask:

Can the framework analyze diverse domains?

Longitudinal Reference Cases ask:

Does the framework evolve responsibly as evidence evolves?

Adversarial Reference Cases ask:

Where does the framework break, overreach, overcomplicate, or manufacture certainty?

Current ARCs