Editorial Methodology & the Use of AI
How Blood Niobium researches, sources, verifies and publishes — and exactly how, and how far, artificial intelligence is used in that process. Transparency about method is part of the method.
01Who We Are — and What We Are Not
Blood Niobium is the first independent intelligence platform dedicated to exposing hidden liabilities — legal, accounting, regulatory, and humanitarian — in critical mineral supply chains. It documents the niobium operations of CMOC Group Limited (HKEX:3993) at the Boa Vista mine in Catalão, Brazil. We do not accuse: we document facts, omissions, obligations and violations, each anchored to a primary or public source. We are not activists and we represent no party to the litigation. Every public piece is factual, dated, and offered to CMOC for right-of-reply before publication.
02Sources & Verification
Every material claim is anchored to a primary or public source:
- Corporate filings — HKEX / SSE disclosures and CMOC Annual Reports (audited by Deloitte).
- Official data — USGS Mineral Commodity Summaries, IBRAM, ANM, Argus Media (named market pricing).
- Court & legal record — TJ-GO proceedings; normative texts (IAS 37, IAS 36, HKEX Rule 13.09, EU CRMA, Germany's LkSG, CSDDD).
Canonical figures — for example, ~US$3.5 billion extracted, the 2024 production record, and R$0 provisioned under IAS 37 over the disputed period — are fixed and never reformulated. Quotations from CMOC are reproduced verbatim, with reference; we do not paraphrase the company in formal pieces. Matters in dispute are described as what the record shows, not as opinion.
03How We Use Artificial Intelligence
We use AI as a production tool, always under human editorial control, for four purposes:
- Translation between English, Portuguese and Simplified Chinese — with canonical terminology protected (e.g., "Material Omission" / 重大遗漏 is never softened to "minor oversight").
- Drafting and structuring assistance for text written from already-sourced material.
- Organising and cross-checking large document sets and datasets.
- Editorial illustrations and diagrams that are clearly illustrative — never presented as photographic evidence.
04What AI Does Not Do — Our Safeguards
05Corrections & Contact
We correct errors of fact promptly and transparently, and corrections are dated. If you identify an inaccuracy — or you are a party with a right of reply — write to info@bloodniobium.org.