1. What this is
New York City Local Law 144 requires every Automated Employment Decision Tool (AEDT) used to screen NYC-located candidates to be audited annually by an independent auditor for selection-rate disparity across sex, race/ethnicity, and intersectional categories. The auditor publishes a public summary; this page hosts ours.
Vettika is the tool. The audit covers the most recent calendar year of recruiter-issued screening interviews where the candidate's self-reported location was NYC. Candidates who declined to disclose demographic data are excluded from the numerator + denominator of the disparate-impact ratio per the law's methodology.
Status notice: Vettika is in early-access and no independent bias audit has been completed yet. We do not yet collect the self-reported demographic data such an audit requires — an opt-in collection step is planned. We will engage an independent auditor once a sufficient NYC-relevant interview cohort exists, and will publish the signed results on this page within 10 business days of receiving them.
2. Methodology
The auditor calculates a selection rate per demographic category as: candidates advanced to the next round divided by candidates screened. The impact ratio is the selection rate of the least-favoured group divided by the selection rate of the most-favoured group, expressed as a fraction of 1.00.
- An impact ratio at or above 0.80 (the “four-fifths rule”) is conventional.
- An impact ratio below 0.80 triggers a follow-up review with model fairness controls.
Categories audited: sex (male / female / unknown), race/ethnicity (Asian / Black / Hispanic-Latino / White / Native / Pacific-Islander / multi-racial / unknown), and the full intersection of the two.
3. Latest summary — no audit completed yet
There are no audit results to report. No independent audit has been performed, no audit window has closed, and no demographic data has been collected. This section will be replaced with the auditor's signed summary when the first audit completes.
When published, the summary will report, per the methodology above: selection rate and impact ratio by sex, by race/ethnicity, and for the full intersection of the two, covering the audited window of NYC-relevant screening interviews.
4. Corrections + flagging issues
If you believe an AEDT decision treated you unfairly, email compliance@vettika.com with the apply link or interview id. We review every complaint individually and escalate repeated patterns to engineering review of the scoring prompts and configuration.
A non-AI alternative is always available — see Request a non-AI alternative.