HomeCountiesArticlesMethodologyContactCAERI is a reference index that communities, researchers, and institutions cite and rely on. This policy states, in plain terms, how the index is governed and how and when it changes.
Version 1.0 · Effective August 8, 2026 · Applies to: Community AI Exposure & Resilience Index (CAERI)
CAERI (pronounced “CARE-ee”) is a reference index that communities, researchers, and institutions cite and rely on. This policy states, in plain terms, how the index is governed and how and when it changes — so that anyone using a CAERI figure knows the rules under which that figure was produced and may be revised. It is modeled on the transparency and accountability principles that credible index providers follow (in the spirit of the IOSCO Principles for Financial Benchmarks), scaled honestly to how StrataHelm actually operates.
CAERI is produced and maintained by StrataHelm. There is currently no external governance board; the integrity of the index rests on disciplines that are enforced in our process and, where possible, in code:
Should an external advisory panel be established in future, this section will be updated to describe it.
CAERI uses semantic versioning (MAJOR.MINOR.PATCH):
| Level | Meaning | Example |
|---|---|---|
| PATCH | A data refresh — new federal data vintages, same methodology. Scores may move; the method does not. | quarterly/annual source updates |
| MINOR | A recalibration — pillar/ensemble weights or thresholds are re-derived and re-validated. Methodology structure unchanged. | annual weight review |
| MAJOR | A methodology revision — the structure, pillars, or formulas change. | a new pillar; a changed scoring model |
A separate correction track covers errors (see §7). Every change is recorded in the public changelog with its effective date; prior citations resolve by canonical URL, which never changes.
CAERI measures exposure to a technology that is advancing unusually fast. The capabilities that define “AI-driven labor disruption” today may look materially different within a year, and the peer-reviewed evidence base is moving with it.
StrataHelm therefore reserves the right to recalibrate the index outside the normal annual cadence when, in our judgment, doing so is necessary to keep CAERI a faithful and relevant measure of current conditions — for example, following a significant shift in AI capability or adoption, or a substantive change in the research on occupational exposure.
Any such off-cycle recalibration is held to the same standards as a scheduled one: it is versioned, sensitivity-tested, documented in the changelog with its rationale and effective date, and prior versions remain reproducible. Off-cycle recalibrations are the exception, not the routine; the scheduled annual review remains the default, and we will not use this provision to make frequent or opaque changes.
caeri: identifier scheme, per-county data
endpoints, and license terms are permanent public commitments and are not
broken by any version change.If we discover an error that materially affects published scores, we will: correct it promptly; record the correction in the changelog as a distinct entry (what was wrong, what it affected, when it was fixed); and preserve the prior, version-stamped output for audit. We will not quietly overwrite a bad number.
This is a governance policy, not the methodology. The specific pillar and ensemble weights, estimation methods, and internal thresholds are proprietary and are not published here. Their existence, governance, and change process are covered by this policy; their values are not. The public methodology is here.
Questions about this policy or a specific index version: contact us.