Validation status, check by check.
Section 11.7 of the Standard sets eight validation checks, V1 to V8. A version of the Standard and its official instrument is validated when the current validation report shows V1 to V5. V6, V7 and V8 ground the prediction and fairness claims in section 2.1. This page says what each check is and where it stands today.
Status on 18 October 2026: no validation report has been published.
What this means today
No result may be described as validated yet.
Section 2.1 of the Standard says a result or credential is validated when it meets checks V1 to V5, as published in the current validation report. The hub of the Standard shows that row exactly as the Standard writes it. No validation report has been published yet, and check V5 needs real assessments, which have not started because no sitting has run.
So, today, no result or credential under AIRS-J/1 may be described as validated, psychometric, predictive of job performance or bias-free. Each of those words rests on checks that are not yet published (section 2.1). A holder or organisation that states a claim beyond its basis in the Standard is acting outside it.
What a result does carry today is its basis for employment use: AIR APAC's expert reading of the person's judgment, given only from an employment-grade result, with every level traced to the person's own acts and a published rule (section 2.1, sections 8.8 and 11.4).
The eight checks
V1 to V8, and where each one stands.
The check column quotes section 11.7 of AIRS-J/1 v2.1; V8 is shortened, and its full text is in the Standard. A result from synthetic candidates is labelled synthetic, as the Standard requires.
| Check | What it checks | Status on 18 October 2026 |
|---|---|---|
| V1 No passing by ritual | Blanket strategies (always accept, always challenge, escalate everything, decide at once, never verify) score below a calibrated reference strategy on every scenario, and none reaches Level 2 on any dimension it games. | Synthetic. The first cut scores were set using the synthetic panels in checks V1 to V3 (section 11.1). StepSim's automated tests include a check that blanket strategies do not reach Level 2 on the dimension they game. No result is published yet. |
| V2 Known profiles land where they should | Synthetic candidates and persona panels with declared decision styles reach the levels their styles map to, at a rate declared before the run. | Synthetic. Used to set the first cut scores (section 11.1). The declared rate and the result are not published yet. |
| V3 Stable across scenarios | The same candidate profile reaches the same level, or one step from it, across parallel scenarios. | Synthetic. Used to set the first cut scores (section 11.1). Until there are real candidates under V5, the report gives this figure from synthetic panels, labelled as synthetic (section 11.8). Not published yet. |
| V4 Readings agree | Every case reading behind a rule (which advice is sound, which evidence matters) is rated independently and agreed, with agreement recorded. | Synthetic working record. In October 2026, two AI raters, working blind and separately, labelled each piece of advice in one StepSim scenario as sound or flawed, over five rounds. The agreement is not yet published in a validation report. |
| V5 Real assessments | V2 and V3 are re-run on real candidates as assessments accumulate, and on agreement between the instrument's level and an authorised facilitator's independent reading. The report states the count. | Not started. No assessment sitting has run yet, so the count of real assessments is zero. |
| V6 Prediction | A level from three scenarios goes with better outcomes on an unseen fourth, by a margin declared before the run. Later performance at work is check V8. | No run is published yet. |
| V7 Fairness | Matched candidates who differ only in gender, age, nationality, first language, market or job level reach the same levels within a declared tolerance; real outcomes are monitored under Section 15. | No result is published yet. Monitoring of real outcomes needs real assessments. The first fairness summary is due in 2027 (section 15). |
| V8 Job performance | In a study of real people in a named role family, Assessment results go with later rated performance at work, by a margin and on a criterion declared before the run. V8 uses real people only, never synthetic candidates. The report is published whether the result is favourable or not, and a result for one role family supports no claim for another. The external academic reviewer or an independent psychometrician checks the analysis before publication. | No study has run. A study needs Assessment results from real people, and no sitting has run yet. The Standard plans the first study with one employer or recruiter as criterion partner (Annex A, item A19). Until V8 is published for a role family, the claim for that role family is the basis "For hiring, promotion and employee assessment" in section 2.1. |
The validation report
When the Standard requires a report, and what it must carry.
- The validation report is versioned with the Standard and published on airapac.org (section 11.7). The evidence programme requires V1 to V8 to be reported for each version (Annex A, item A14). The Standard sets no calendar date for the first report.
- The report labels every synthetic result as synthetic. Synthetic candidates may be used for V1 to V7 when their personas carry the demographic and role detail of the people they stand for. V8 uses real people only (section 11.7).
- The report carries an annual impartiality statement: the share of certified candidates who trained with AIR APAC, AIR Academy or a licensed trainer before certifying, and any other overlap between training and certification (section 11.7).
- The report states the measurement precision for each track and role edition: agreement across parallel forms, the standard error of measurement at the cut and at the floors, and classification consistency at the cut (section 11.8).
- Where monitoring shows a claim needs strengthening, AIR APAC fixes the instrument, the rule or the anchor in the next release, and the validation report records what changed (section 4.8).
Governance: the review board and the external academic reviewer