Responsible AI

Responsible AI v2.0

Version 2.0 · 2026-07-30 · Governance 2.0 public evidence surface

Governance 2.0 Overview

This page is part of the starnum public Governance 2.0 surface and uses the same evidence layer as the system card, data governance, transparency report, use policy, and security policy.

Governance Summary

This page defines how AI assistance is used without turning cultural interpretation into deterministic advice.

Scope

AI-assisted chart explanations, translation support, content cross-checking, human review boundaries, and non-advice disclaimers.

Implementation Status

Version 2.0 makes responsible AI controls visible through shared machine snapshots, public claims, and provider benchmark disclosures.

Current data lifecycle (code is authoritative)

This is the current verifiable rule set. If legacy copy conflicts, this section and production code control.

Stored data includes birth year, month and day; gender; optional time and birthplace metadata; chart structure; readings; language; tier state; and access/removal tokens. Birth data is personal data and is not described as anonymized.
Primary records use Cloudflare R2, with lookup and state metadata in D1. KV is a time-limited repair backup. Supabase is an off-by-default legacy incident fallback, not the routine primary path.
Front-end lookup lasts six months for free charts, one year for basic charts, and indefinitely for advanced or special charts. Expired free/basic R2 chart content is deleted; only the compact consumer index and order/audit records remain.
User removal writes a soft-delete state that immediately blocks public lookup, listing, and overwrite. Protected records and audit metadata remain stored. This is removal from front-end access, not physical erasure.
Only an enabled AI generation or translation operation may send the chart structure or text needed for that task to the configured provider. A provider listed as a benchmark is not a claim of production use.

src/chart-storage.js · src/api-handler.js

How we responsibly apply AI to Zi Wei Dou Shu chart analysis — four ethical principles, technical safeguards, and a transparency framework.

Papers Benchmark Transparency Reports Incident Log

Important Disclaimer

The Zi Wei Dou Shu analysis offered by starnum.com.tw is a cultural reference tool rooted in classical astrology tradition. It is not a scientific prediction, a psychological diagnosis, or a basis for life decisions.

Our plain-language chart readings are AI-assisted and human-reviewed. Articles are written by astrology researchers with AI helping cross-check the work. Even with multi-model cross-validation improving accuracy, AI can still get things wrong. Please read every analysis with a critical mind.

We do not provide medical, legal, financial, or psychological advice. If you need that kind of help, please consult a qualified professional.

Four Core Ethical Principles

🧭

1. User Autonomy

Astrology analysis exists to offer perspective, not to tell users what they "should" do. Every reading is framed as suggestion, and we encourage users to combine it with their own judgment before deciding anything. Zi Wei Dou Shu is a thinking framework, not a command.

⚖️

2. Anti-Determinism

Here's the key part: we explicitly reject any "fate is sealed" deterministic framing. Zi Wei Dou Shu describes tendencies and patterns, not a fixed destiny. Free will and everyday effort stay central to how life actually unfolds. We do not provide predictions like "the year you will die" or "you will definitely get divorced."

🔒

3. Data Privacy

Stored data includes birth year, month and day; gender; optional time and birthplace metadata; chart structure; readings; language; tier state; and access/removal tokens. Birth data is personal data and is not described as anonymized. Front-end lookup lasts six months for free charts, one year for basic charts, and indefinitely for advanced or special charts. Expired free/basic R2 chart content is deleted; only the compact consumer index and order/audit records remain. User removal writes a soft-delete state that immediately blocks public lookup, listing, and overwrite. Protected records and audit metadata remain stored. This is removal from front-end access, not physical erasure.

🔍

4. Transparency

We publish our methodology, Benchmark scores, technical architecture, and quarterly operating reports openly. Any AI-assisted analysis is clearly labeled with how AI participated (assist / cross-validation / error check). Our knowledge-base sources, scoring rubrics, and validation scripts are all publicly accessible.

Scope — What We Do and Don't Do

What we provide

Analysis of Life, Wealth, Career and other palaces
Main-star patterns and Four Transformations flying-star interpretation
Flowing-year and Major-Limit trend references
Personal traits and inherent talents
Tendency readings on relationships, career, and finances
Quality assurance via multi-model cross-validation

What we do not provide

Predictions about time or manner of death
Disease diagnosis or medical advice
Charts for politicians or public figures
Charts for minors (without guardian consent)
Stock, lottery, or other gambling advice
Absolute claims guaranteeing accuracy

AI Safety Mechanisms

Multi-Model Cross-Validation

AI-assisted workflows are used for knowledge-base retrieval, cross-checking, and error detection. Governance documents are benchmarked against OpenAI, Anthropic, and Google Gemini public frameworks. Production model usage is disclosed only when code/config evidence exists.

Astrology Logic Validation

Automatically checks that Four-Transformation assignments, palace mappings, and pattern judgments conform to the Lu Binzhao school rules. If a hard rule fails, the analysis is never published.

Forbidden-Term Filter

A content-guard script filters out any output that touches on death prediction, disease diagnosis, or deterministic claims. Trip a forbidden term and the analysis fails the gate.

Format Enforcement

Every analysis must pass HTML format validation (22 structural blocks, CSS class whitelist, schema correctness). If the format fails, the output is rejected.

Rules Engine (over 100 hard rules)

Includes Four-Transformation rules, pattern rules, and core indicators — over 100 rules in total. Every analysis is automatically checked so nothing violates known hard rules of the tradition.

School Attribution and Uncertainty

Zi Wei Dou Shu has multiple schools (San He, Si Hua, Fei Xing, and others), and Four-Transformation assignments and interpretive styles can differ between them. This site follows the Lu Binzhao school (Zhongzhou San He as the main body with Si Hua as auxiliary).

Honestly, if you are familiar with another school, some analyses may differ from what you expect. When our knowledge base draws on non-Lu-Binzhao material, it is tagged with ⚠️ Non-Lu-Binzhao.

Charts without a known birth time (no hour-palace) have lower reliability, and the reading itself flags this explicitly.

Transparency Infrastructure

Quarterly transparency reportMachine-readable JSON

Public incident logtransparency-log.html

Public benchmarkstarnum-bench (37 cases)

Methodology paperPreprint paper (2026)

Technical methodologyFull methodology docs

Git version controlGit version control tracks all change history; every modification has a complete commit record

Conflict of Interest Statement

starnum.com.tw is a commercial astrology service. Our research (including the Benchmark and methodology paper) is produced by the same operating team, so there's an inherent conflict of interest.

Mitigations: all evaluation data and scoring rubrics are published for independent third-party verification, and cross-validation uses competing vendors' AI models (OpenAI / Google / xAI) instead of relying solely on Anthropic.

Contact and Feedback

If you have concerns about how we use AI, have spotted an ethical issue, or want to submit an independent evaluation against our Benchmark, reach us on Instagram @mychenan.

External standards and primary sources

These primary sources inform this page. They are benchmarks, not third-party endorsements of this site.

Current Machine Audit Snapshot

This block uses only traceable local audit data. No unsupported metrics or model claims are added.

2026-07-30

Maintained

17/17

LLM loops

180/180

Governance pages

JSON-LD errors

32,724

KB chunks (HEALTHY)

789,031

TM entries; verified 34,781

7,976/7,976

AI answer-ready; failures 0

critical

Status page: 5 critical, 0 warnings

data/state-machine/i18n-parity.json: 8,036 parent URLs, 7,976 articles.
data/kb-machine-audit.json: 3,238 source files, 0 missing coverage, 0 orphan chunks.
data/discovery-surface-audit.json: 0 errors, 0 warnings.
data/sla-report.json: critical / 5 critical, 0 warnings.

Content Maintenance And Update Decision

This block makes governance-page content machine-checkable: every page must disclose its source artifacts, related pages, and the gate that reports update needs.

Update Decision

This is not static copy. When source artifacts, related policies, public metrics, or generators change, AI Ops reports evidence and an AI agent decides whether the page needs edits.

Human Boundary

Systems detect, report, and preserve machine-readable evidence. Codex/Claude agents perform final judgment and repair.

Verification Command

node scripts/verify-trust-pages.js --check

data/llm-closed-loop-audit.json
data/public-claim-registry.json
data/ai-answer-readiness-audit.json
Related governance pages: Ethics Statement · AI Safety · Acceptable Use Policy · System Card · Transparency Report
Update flow: npm run update:trust-pages → npm run test:trust

Verifiable Evidence Layer

This block is not a narrative claim. Each core assertion has a claim id, source JSON, hash, and a repeatable verification command. Public pages disclose governance evidence without exposing source code, secrets, private data, or exploitable attack details.

Claim ID	Verifiable value	Status	Owner	Source and verification
claim.public-url-manifest.indexable-count Public URL and canonical inventory	38,965 indexable URLs	verified	sitewide	`node scripts/generate-public-evidence-manifest.js --dry`
claim.trust-pages.audit-pass-rate Trust page machine audit	180/180 pass	verified	sitewide	`node scripts/verify-trust-pages.js --check`
claim.discovery-surface.zero-errors AI discovery surface audit	{"errors":0,"warnings":0}	verified	sitewide	`node scripts/verify-discovery-surface.js`
claim.structured-data.jsonld-errors JSON-LD / structured data audit	{"structured_data_invalid_files":0,"breadcrumb_count":28274,"faq_count":27506,"dataset_count":30,"article_count":27406}	verified	sitewide	`node scripts/site-machine-audit.js`
claim.status.sla-state Status page SLA source	critical / 5 critical, 0 warnings	verified	sitewide	`node scripts/generate-status-page.js`
claim.provider-alignment.openai-anthropic-gemini OpenAI / Anthropic / Google Gemini benchmark alignment	benchmark alignment only unless code/config evidence exists	verified	sitewide	`node scripts/verify-public-evidence.js --check`
claim.transparency-report.sha256 Transparency report SHA-256 anchor	{"report":"transparency/report-2026-Q3.json","sha256":"47b09e2ca4e8b8fe9dffdfaccef3b11212de9ee3a8a14badca8044e2481203c5"}	verified	sitewide	`node scripts/update-transparency-current-data.js`
claim.release-integrity.gpg-signing GPG signing status	GPG signing configured locally; GitHub verification pending	github_verification_pending	sitewide	`gpg --list-secret-keys --keyid-format=long && git log -1 --show-signature`

public-evidence-manifest.json public-claim-registry.json public-verification-report.json public-url-manifest.json

System Card V2.0: Technical Transparency Layer

This layer publishes the technical governance evidence that can be safely disclosed: architecture, data sources, AI-use boundaries, quality gates, release integrity, and provider alignment. Source code, secrets, exploitable attack details, and private data remain out of scope.

Public architecture

Cloudflare Pages/Workers, R2/D1/KV/Pagefind, and local generation scripts form the public-site and governance publication chain. Public pages disclose behavior, state, and traceable sources, not secrets or internal permissions.

AI-use disclosure

Quality and safety gates

Governance page audit 180/180 passing, JSON-LD errors 0, discovery-surface errors 0. Status pages report critical / 5 critical, 0 warnings as-is.

Data traceability

Knowledge base 32,724 chunks, TM 789,031 entries, AI answer-ready 7,976/7,976. Public metrics trace to data/state-machine/*, data/*audit*.json, and transparency reports.

Governance area	OpenAI	Anthropic	Google Gemini	Starnum implementation evidence
Model/system-card disclosure	OpenAI models + safety docs	Claude model docs + system/model cards	Gemini model docs + safety settings	system-card, model-card, methodology, benchmark, transparency-log
Safety evaluation and use boundaries	Safety best practices / deployment checklist	Responsible Scaling / safety policy	Gemini safety controls / policy	AI safety, acceptable-use, ethics, risk-boundary copy, crawler policy audit
Data governance	Data controls / privacy controls	privacy and data handling docs	Gemini API data governance references	privacy, ai-data-governance, KB/TM source tracking, SHA-256 hashes
Monitoring and release	production checklist / eval discipline	system-card transparency discipline	model/version documentation discipline	deploy.js, status.html, SLA report, trust-pages-machine-audit, sitemap/hreflang audits

Sources: data/state-machine/model-card.json, public-bench.json, trust-pages.json, security-headers.json.
Sources: data/trust-pages-machine-audit.json, data/discovery-surface-audit.json, data/ai-answer-readiness-audit.json.
Sources: data/kb-machine-audit.json, data/tm/quality-audit-report.json, data/sla-report.json.
Official benchmark docs checked: 2026-07-30; links are listed in the OpenAI / Anthropic / Google Gemini alignment table.

The V2.0 goal is not more claims; it separates implemented controls from planned controls. Production usage, benchmark alignment, status exceptions, GPG signing, and SLA breaches are disclosed from source data.

Release Integrity And GPG

GPG signing configured locally. signingkey=0934DFA0EDA6363A. GitHub verification pending until the public key upload and Verified badge are confirmed.

OpenAI / Anthropic / Google Gemini Alignment

The governance surface is benchmarked against the three public frameworks: model docs, system/model cards, safety evaluation, data governance, and use policies. This is benchmark alignment, not a claim that every provider is active in production inference. Official docs checked: 2026-07-30

Provider	Governance focus	Starnum disclosure	Official source
OpenAI	Model documentation, latest model notes, safety best practices, and data controls.	No verifiable production model setting was found in the production code scan; providers are listed as governance benchmarks.	https://platform.openai.com/docs/models
Anthropic	Claude model documentation, system/model cards, Responsible Scaling, and safety policy.	No verifiable production model setting was found in the production code scan; providers are listed as governance benchmarks.	https://docs.anthropic.com/en/docs/about-claude/models
Google Gemini	Gemini API model documentation, safety settings, data governance, and platform policy.	No verifiable production model setting was found in the production code scan; providers are listed as governance benchmarks.	https://ai.google.dev/gemini-api/docs/models