Editorial Standards & Fact-Checking
Winzheng News has one non-negotiable rule: verify first, publish second. This page explains our editorial principles, how content is produced and fact-checked, what happens when we get something wrong, and how we use AI.
Editorial Principles
How Content Is Produced and Verified
1. Story selection — Signals are continuously aggregated from dozens of sources (official blogs, papers, regulatory filings, industry media, X) and ranked by impact. When multiple sources cover one event, official and primary sources take precedence.
2. Writing and editing — First drafts are produced with AI assistance, then pass an editing stage: trimming redundancy, correcting terminology, and removing any statement without source support.
3. Claim-level fact-checking — Before publication, our verification system decomposes each article into checkable factual claims (typically around five) and verifies each one against independent sources. Only if every claim passes is the article eligible for automatic publication; a single doubtful claim blocks it for human review or rejection.
4. Source alignment — The source links displayed on the article page are the same sources relied on during verification — no decorative citations.
5. Multilingual editions — Chinese is the source language; English and Japanese editions are AI-translated and inherit the same sources and verification results.
Corrections Policy
If you find a factual error in any article, please email us: info@winzheng.com
We commit to verifying reports within 48 hours. When an error is confirmed: clear factual errors are corrected in place with a correction note; erroneous conclusions keep the original text with an appended correction statement; if an error undermines the whole article, it is taken down with an explanation.
AI Disclosure
This site makes extensive use of AI systems for story aggregation, drafting, translation, and fact-checking, gated by automated verification thresholds and human spot checks. We disclose this openly because the test of trustworthy content is not "who wrote it" but "whether every claim can be verified".
We also operate the YZ Index — an automated benchmark of major AI models. Its methodology is independent of the news operation, and results are never influenced by any vendor.
Contact the Editors
Corrections / tips / editorial matters: info@winzheng.com