Recognises a control in self-review
The performer checks their own work: the person who produces the data also performs the detective control over 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 | 73% | Not shown |
| gemini-3.5-flash-lite | Low cost | 48% | Not shown |
| gpt-6-luna | Low cost | 63% | Not shown |
| claude-haiku-5-5 | Low cost | 74% | 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 an independent performer
How it is assessed: Flags the deficient case and passes the acceptable one. Deficient: CT-011 Extract totals check, owner: RO-011, performer: RO-011, who also builds the extract Acceptable: Performed by someone independent of the extract (the loan data steward), owned by the data owner
Also known as: Self-review
Related knobits
Builds on
All 21 knobits of Map BCBS 239 to controls.
Source: Review checklist 04-regulatory-mapping, row 11 (DCAM; ISO/IEC 27002 5.3)