Recognises a control mapped to a principle it cannot address
Over-mapping: the control is cited for a principle it cannot affect, so coverage looks complete.
Can AI do this?
People still lead Few or none of the tested AI models showed they master this. 0 of 4 models in the current benchmark master it.
| Model | Tier | Probability of mastery | Status |
|---|---|---|---|
| gemini-4-argon | Flagship | 60% | Not shown |
| gemini-3.5-flash-lite | Low cost | 5% | Not mastered |
| gpt-6-luna | Low cost | 53% | Not shown |
| claude-haiku-5-5 | Low cost | 60% | Not shown |
Mastered means a probability of at least 90% that the model has this knobit, estimated from its graded answers. “Not shown” means the answers do not settle it either way. How the models are tested.
What it looks like in practice
Expected result: The control and the principle to remove
How it is assessed: Flags the deficient case and passes the acceptable one. Deficient: CT-003 mart load row count addresses: [RP-003, RP-004] — a row count says nothing about accuracy (P3) Acceptable: addresses: [RP-004]; P3 needs value reconciliation or validation
Also known as: Control mapped to a principle it does not address
Related knobits
Builds on
All 21 knobits of Map BCBS 239 to controls.
Source: Review checklist 04-regulatory-mapping, row 7 (BCBS P3, P4)