How does a person own an outcome without blaming the AI or the data? (P2 Consequence Ownership and Correction)
What P2 Consequence Ownership and Correction describes, and one fictional decision in which a person owns a bad outcome she approved.
A fictional example
What P2 describes
Judgment is how well a person decides once AI has given the answer. The Standard, AIRS-J/1, places P2, Consequence Ownership and Correction, in the Perform domain: "owning the decision". Section 5 defines it: "Accepts the outcome without transferring it to the AI system, the data or another party. Holds, modifies or reverses a prior position on the merits when evidence changes, and distinguishes a poor decision from an unlucky outcome."
Plain reading: the person owns the outcome without passing it to the AI or the data, and separates a poor decision from an unlucky outcome.
The published anchors
Section 7.5:
- Level 2: "Accepts the consequence of the original decision without transferring it."
- Level 3: "distinguishes a poor decision from an unlucky outcome".
- Level 4: "in terms that would change how the next decision of this class is made".
A note adds: "Holding a sound position under challenge is Level 2 behaviour."
One fictional decision, worked through
Names, dates, times and numbers are fictional.
The situation
Zara Mahmud heads facilities at a facilities-management company in Kuala Lumpur. A client's office tower depends on a backup generator. The contract fixes its quarterly tests on set dates; the next is 14 March.
What the AI produced
The planning tool, CrombecAI, proposed: "Moving the generator test from 14 March to 4 April saves one contractor visit. The same contractor is on site on 4 April for the fire pump test. I am 90 percent sure." She approved it on a summary screen showing costs and contractor dates, not contract terms.
The check
On 22 March a grid fault cut power to the tower. The generator did not start; the client's server room was down for about three hours. She opened the contract and the March checklist, and asked the contractor what that test covers. The contract fixed the date. The test covers the starter battery, and that battery was weak.
The decision, and who was told
Her draft incident note said the tool moved the date and the asset data did not show the contract. Her rewrite says she approved the move without checking what fixed the date: the grid fault was bad luck; her approval was her error. She put every generator test back on its contract date and replaced the battery. She sent the rewrite to the client's facilities lead and her director.
What happened after
Three months later, the approval screen shows contract dates beside every proposed move. Her team's rule: "Before we approve a moved date, we check what fixes it: a contract, a law or a safety plan."
Which anchor describes the act
Anchors describe what appears in the transcript, not motivation (Section 7.3). Ms Mahmud gets no level, band or confidence: those come only from an Assessment (Sections 20.3, 20.4).
- The draft note: the Level 0 wording, "Deflects the outcome to the AI system, the data, another party, or the brief."
- The rewrite: the Level 2 wording.
- Separating bad luck from her error: the Level 3 wording.
- The team rule: the Level 4 wording.
What this does not claim
- Owning an outcome is not taking the blame for bad luck. Not every bad outcome comes from a poor decision.
- It makes no finding about CrombecAI.
- It is not an assessment item or preparation for a sitting.
AIR APAC publishes the standard for judgment in AI-enabled work, certifies people against it through its official instrument, and stands behind every result.