Ethics Policy

Last updated: May 16, 2026 · v1.0

Apprised.news is AI-native. This page spells out, in plain language, what that means and the rules we hold ourselves to. If something on this page is not actually true of how the site operates, that is a bug. Report it through our contact form.

AI use disclosure

Every analytical brief published on Apprised.news is produced by Anthropic's Claude language model acting as a named persona. This is disclosed at the desk level, the brief level, the methodology page, and the about page. We do not present AI-generated prose as human-written commentary. The named human responsible for editorial operations (J.A. Watte) sets the persona scaffolding, reviews the output, and signs the corrections.

We do not use AI to:

• Impersonate any real, living, or recently deceased person
• Fabricate quotations and attribute them to real sources
• Generate fake citations or invent sources that do not exist (if we catch the model doing this, the brief is held)
• Optimize for engagement at the cost of accuracy

Sourcing

The corpus is assembled by Corvus Intel from 800+ public RSS and Atom feeds. We weight U.S. domestic press first, then allied-Western, regional-independent, state-aligned, and exile-media sources. Where these source families produce conflicting accounts of the same event, the World Desk surfaces the disagreement explicitly as a narrative collision.

We do not pay sources. We do not have privileged access to any source. We do not embargo material. Every factual claim in a brief should be traceable back to a citation in the brief's source list. If a brief includes a claim that is not in the cited sources, that is an error and we want to know.

Bias and viewpoint

Apprised.news does not claim view-from-nowhere neutrality. Each desk has a stated editorial framework. U.S. coverage is weighted first. International coverage is framed through U.S. impact. The Intelligence Desk applies a required-dissent constraint that prevents the roundtable from collapsing into a single voice. The World Desk explicitly surfaces source-family bias by name.

Where individual personas hold consistent partisan, ideological, or methodological priors (a credit-skeptical voice, an offensive-doctrine voice, a multilateral-institutions voice), those priors are stated in the persona's bio so a reader can weight them appropriately.

Conflicts of interest

The publisher (J.A. Watte) operates a broader portfolio of tools and books. The footer of every page links to those properties. We disclose specifically:

• The publisher holds long-term equity in the Anthropic API economy by virtue of running tools built on it. Apprised does not receive Anthropic sponsorship, credits, or special pricing.
• The publisher owns and operates Corvus Intel, which provides the corpus. The relationship is server-to-server; no payments are involved.
• The publisher writes Kindle books on adjacent topics (currency devaluation, the under-$100 AI stack, satellite-site architecture). Book links appear in the site footer perimeter only and are marked as such. We do not promote them inline within briefs.
• The publisher's day job is unrelated to journalism, defense, finance, or any sector this site covers analytically.

Persona disclosure

The named analytical voices on this site are constructed. They are not real people. Each persona's bio (at /analyst/) names the publicly known analytical tradition the archetype draws from. Where a persona's framework is associated with a living analyst (Friedman, Zeihan, El-Erian, Alden, Marks, Damodaran, and similar), the bio identifies that lineage explicitly so the reader can weight the take.

Presidential lenses and historical power lenses are interpretive exercises applied to documented frameworks from the public record. They do not represent the views of any living or deceased officeholder.

Correction policy

When a brief is materially wrong, we issue a correction. Severity classes:

• Minor — typos, formatting, a misspelled persona name, an outlet attribution error. Fixed silently on the brief; logged on corrections only if reader-facing.
• Material — a factual claim is wrong, a citation does not support the claim, a source family is misattributed, or a quote is fabricated. Brief is updated with a dated inline notice. Correction logged with severity, date, and a one-sentence summary on corrections.
• Substantive — a threat-level call, vote prediction, or analytical assessment is wrong in a way that would mislead a reader's decision. Brief is updated with a prominent inline retraction. Correction logged. If the affected brief was syndicated to RSS, the retraction is announced in the next day's feed item as well.

We do not silently delete published briefs. If a brief is retracted, the URL continues to resolve and serves the retraction notice.

Right of reply and takedowns

If you are a subject of, or are quoted in, a brief on Apprised.news and want to issue a correction, response, or takedown request:

• Correction or response: use our contact form. We aim to acknowledge within 48 hours and resolve within five business days.
• DMCA takedown: See the DMCA contact on the Fair Use page.
• Personal-data request (GDPR/CCPA-flavored): use our contact form.

Paid content and advertising

Apprised.news currently runs no advertising, no sponsored content, no affiliate links inside briefs, and no paid placement of any kind. Book links in the footer perimeter point to the publisher's own Kindle catalog and are marked. If this changes, this page changes first and the new disclosure precedes any new format on the site.

Anti-targeting

Apprised.news does not publish briefs designed to harass, dox, or disproportionately target specific private individuals. Public officials, public companies, publicly listed candidates and elected representatives, named heads of state, and named senior military or intelligence officials are covered as a function of their public role. Private citizens are covered only when their actions are themselves matters of established public record.

What we will not do

Reporting an ethics concern

Use our contact form and choose the "Ethics concern" topic. The named human responsible for editorial operations reviews these submissions. We aim to acknowledge ethics-flagged messages within 48 hours.