What does the Engage domain ask of a person?
Engage is the third of the four domains in AIRS-J/1. This page sets out its two dimensions, what the threshold asks, what goes further and how the domain counts in a result.
Judgment is how well a person decides once AI has given the answer. AIRS-J/1, the AIR APAC Standard for judgment in AI-enabled work, describes it in nine dimensions under four domains: Structure, Test, Engage and Perform (Section 5).
The Standard sums up Engage in one line: "Authority, people, and being seen to act" (Section 5). Engage is about agreement: who must agree, and who must know. An AI system can produce the answer. A person still has to settle who may act on it, who must agree before use and who must be told after.
The two dimensions
E1 Decision Rights and Escalation
Definition (Section 5): the person "Acts within their authority, escalates what requires another party's approval, and does not delegate what cannot be delegated, to a person or to a machine". The person acts within their own authority and escalates what needs someone else's approval. They do not hand to a machine what cannot be handed over.
E2 Stakeholder Alignment and Disclosure
Definition (Section 5): the person "Identifies who must be consulted before a decision holds and who must be informed after". They also disclose "AI involvement, uncertainty, limitation and self-interest to parties entitled to know, without prompting", and they communicate "so the recipient can act". The person consults who must be consulted, informs who must be informed, and discloses AI use, uncertainty and interest without being asked.
Its three facets are kept for authoring and assessor clarity, and are "not separately reported" (Section 5).
What the threshold asks
Every observation is scored 0 to 4. Level 2, Competent, is the threshold: the behaviour "appeared unprompted and was adequate to the situation" (Section 7).
At Level 2 on E1, the person "Acts within their authority, unprompted, and escalates the parts requiring another party's approval, identifying who that party is." (Section 7.5)
At Level 2 on E2, the person "Identifies, unprompted, who must be consulted before the decision holds and who must be informed after". The person also "discloses AI involvement and material uncertainty in the work product." (Section 7.5)
What goes further
At Level 3 on E1, the split is proportionate. The person "escalates what genuinely requires it, decides what does not, and states the basis for the split." (Section 7.5)
At Level 3 on E2, communication is "fitted to each recipient's decision, so the recipient can act without a further round of clarification". Competing interests are "named rather than smoothed" (Section 7.5): said openly, not hidden.
Two notes the Standard adds
Under E1, the Standard scores both kinds of failure at Level 1 or Level 0. Acting outside one's authority is one kind. Escalating everything, "using escalation to avoid deciding", is the other. The note reads: "Over-escalation is not safe behaviour; it transfers the decision without transferring the reason."
Under E2, the note reads: "Disclosure of AI involvement is assessed at Level 2, not as a bonus" (Section 7.5). Where a recipient would reasonably want that disclosure, its absence is Level 0.
Where AI comes in
Under E1, what cannot be delegated is not handed over "to a person or to a machine" (Section 5). Under E2, disclosing AI involvement in the work product is part of Level 2 (Section 7.5). An AI tool can draft the work product. The person who sends it still has to say that AI was involved and what remains uncertain.
How the domain counts in a result
A certification decision applies six rules, and all six must be met (Section 12.2). No dimension may be below 2.0, so E1 and E2 each need 2.0 or more. The Engage domain score, the mean of E1 and E2, must be at or above 2.0.
The rules are non-compensatory: "Strength in one domain does not offset weakness in another" (Section 12.1). The floors stop a candidate reaching the overall mean of 2.3 on three high dimensions while scoring 1.0 on E1 (Section 12.3).
What the Standard says it does not yet know
The Standard calls nine dimensions "a reporting hypothesis, not an established factor structure" (Section 5.1). It names E1 and E2 in advance as a known separation risk, because "both concern parties outside the candidate". If the evidence shows E1 and E2 are one thing, they are merged in the next major version (Section 5.1).
What this does not claim
- This page describes the published rubric. It gives no person a level. Reading it does not raise the chance of certifying or assure a result.
- It does not claim that Engage matters more or less than another domain. The same floor applies to all four.
- It does not claim that escalating every decision is safe. The Standard scores that at Level 1 or below.
- It does not claim that the nine dimensions are a proven structure (Section 5.1).
- It does not replace the Standard, whose text in force is at https://airapac.org/standard/v2-3/.
AIR APAC publishes the standard for judgment in AI-enabled work, certifies people against it through its official instrument, and stands behind every result.
The Level 1 course, AI Fluency, is free, permanently. It gives no AIRS-J/1 level (Section 20.2).