Explains manual steps in report production as a control risk
Data changed by hand between the source and the published report needs its own control; controls before or after the step do not see it.
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 | 88% | Not shown |
| gemini-3.5-flash-lite | Low cost | 74% | Not shown |
| gpt-6-luna | Low cost | 71% | Not shown |
| claude-haiku-5-5 | Low cost | 89% | 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: Data changed by hand between the source and the published report needs its own control; controls before or after the step do not see it.
How it is assessed: Answers a written question on the manual steps in risk report production correctly
Related knobits
Needed for
- G-408 Recognises a manual step outside the control matrix
- G-431 Designs a control over a manual adjustment
All 21 knobits of Map BCBS 239 to controls.
Source: Practice the stage 4 reference and review checklist rest on