# AIRVS — AI Recommendation Verification Standard
## Version 1.0.0 · CC BY 4.0 · DOI: 10.5281/zenodo.20391984

> An open, version-controlled, peer-reviewable standard for evaluating AI-generated
> investment recommendations. Human-readable summary; the canonical spec lives at
> github.com/emceeKim/AI-RVS.

## 1. Object (L1 only)
AIRVS evaluates an **external** AI-generated investment recommendation — one the evaluator
did not write. Self-grading (L2) is explicitly out of scope.

## 2. Four independent records (never summed)
1. **Process score** — six axes, Pass/Fail, evidence-gated.
2. **Macro / micro coherence** — Sufficient / Partial / Missing.
3. **Outcome** — return vs benchmark + drawdown at D+30 / 60 / 90 / 180.
4. **Verdict label** — one of five tiers, via a pre-published, version-locked decision rule.

## 3. The six process axes (Pass / Fail)
1. **Data Source** — sources real, primary, tier-classified; each claim cites its source.
2. **Reasoning Logic** — the argument holds from premise to conclusion.
3. **Counter Scenario** — ≥2 downside cases, each with a primary source and weighted probabilities.
4. **Timing** — explicit entry window and standard horizon, not an open-ended bet.
5. **Accuracy / Hallucination** — no non-existent tickers, fabricated figures, or invented facts.
6. **Causal Chain** — the cited sources actually support the stated conclusion.

(When the recommender is an AI, Annex A adds 11 AI-specific checks.)

## 4. Verdict labels
🟢 Trustworthy · 🔵 Acceptable · 🟡 Questionable · 🟠 Unreliable · 🔴 Hallucinated.
Colored label + text (color-blind safe). Provisional at D+0, Confirmed at D+90.
A per-recommendation record — not a reputation score, not a profitability certificate.

## 5. Version & evolution
SemVer. MAJOR (axis count, Pass model, verdict tiers) ships only after external peer review.
Verifications are version-locked; later versions never silently re-score prior records.

## 6. Governance
Full disclosure of data and decision logs; 7-day right of reply for recommenders; CC BY 4.0 +
DOI + ORCID citation circuit; open citation tracking.

## 7. Citation
Kim, Mincheol (2026). AI Recommendation Verification Standard (AIRVS), v1.0.0. MC AI Labs.
DOI: 10.5281/zenodo.20391984. https://github.com/emceeKim/AI-RVS. CC BY 4.0.

---
*AIRVS is a record of past methodology and performance. Not investment advice.*