Designs a control over a manual adjustment
A control that logs, approves independently and reconciles the adjustments, mapped to accuracy
Can AI do this?
Not settled yet The benchmark tested this, but the models' answers do not settle yet whether they master it. 0 of 4 models in the current benchmark master it.
| Model | Tier | Probability of mastery | Status |
|---|---|---|---|
| gemini-4-argon | Flagship | 64% | Not shown |
| gemini-3.5-flash-lite | Low cost | 17% | Not shown |
| gpt-6-luna | Low cost | 15% | Not shown |
| claude-haiku-5-5 | Low cost | 67% | 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: A control that logs, approves independently and reconciles the adjustments, mapped to accuracy
How it is assessed: The control is performed by someone other than the adjuster, maps to P3 or P7, and explains why neighbouring controls miss the step
Related knobits
Builds on
- G-454 Explains manual steps in report production as a control risk
- G-452 Explains design and operating effectiveness of controls
All 21 knobits of Map BCBS 239 to controls.
Source: Graded on its own by dg-bcbs239-mapping