Vastkind uses only public sources or named, attributable sources whose identity and claims can be checked. It does not use anonymous, confidential, or untraceable sources.
Acceptable sources
Depending on the claim, acceptable sources can include peer-reviewed research, preprints with their status made clear, official records, legislation and regulatory material, court documents, company filings, technical documentation, datasets, standards, direct statements from identifiable people or organizations, and credible reporting that names or clearly attributes its evidence.
Source hierarchy and independence
Primary evidence is preferred when it is available and intelligible. Secondary reporting can provide context, independent verification, or access to facts that cannot be obtained directly. Company announcements, investor materials, product demonstrations, and sponsored research are treated according to their incentives and limitations. A link does not turn a claim into proof.
Verification and attribution
A source must support the specific claim attached to it. Material factual claims are checked for date, scope, method, and relevant caveats. Sources are attributed clearly enough for readers to identify and examine them. For medical, health, financial, legal, safety, or security-related claims, the standard of evidence and caution is higher.
Generative AI
Generative AI output is never evidence, authority, or a source. It may assist research organization or drafting, but any factual claim it produces must be verified against acceptable sources before publication.
Source changes and corrections
If a source is corrected, withdrawn, materially changed, or contradicted by stronger evidence, Vastkind reassesses the affected claim and updates the article when necessary. Source concerns can be sent to social@vastkind.com.