Purpose
This policy explains how Insider Release may use artificial intelligence (AI) tools in research and drafting, what AI must not do, how we disclose AI use to readers, and who remains responsible for published content. It complements the Editorial Policy and Corrections Policy.
Insider Release’s existing article footers often state that some text and images may be partially or entirely created using AI tools (sometimes naming Grok / xAI), that the site is an informative blog not a registered press site, and that readers should verify critical information independently. This policy upgrades that practice: disclosure should be clear and findable, not only buried in a generic footer—consistent with EU-oriented transparency expectations.
Guiding principles
- Transparency — Readers should be able to tell that AI may assist drafting or imagery, without hunting through fine print alone.
- Human editorial responsibility — A human decides what is published. AI is a tool, not an author and not a publisher.
- Precision over vague slogans — Prefer concrete, accurate disclosures over vague “full AI assistance” or false “human-written only” claims.
- No fabricated evidence — AI output is never an excuse for invented sources, quotations, document numbers, or data.
- Pillar discipline — AI does not expand the site into off-pillar sensational topics; pillars remain Declassified Documents, UAP Disclosure, Systemic Risks.
What AI may do
AI tools may be used to:
- Outline structure and section headings for longform analysis.
- Draft or revise prose for clarity, accessibility, and document-first tone.
- Summarize lengthy public PDFs or transcripts as an aid to editorial work.
- Suggest internal links, FAQ candidates, or schema field drafts for human review.
- Support translation or readability passes (site language is English).
- Assist image generation or editing only when clearly appropriate and disclosed (see Images).
- Help research navigation (e.g. suggesting where a public document might live)—not to invent that the document exists.
What AI must not do
AI tools must not:
- Be credited as an author or co-author.
- Fabricate citations, quotations, FOIA case numbers, URLs, or “leaked” documents.
- Assert personal credentials, degrees, clearances, or awards for Adriano Margarone or others.
- Present speculative UAP, weapons, or AI-capability claims as confirmed fact.
- Generate medical, legal, or operational advice framed as professional counsel.
- Bypass human publish control (no unsupervised auto-publish of AI drafts to the live site).
- Be used to impersonate real people or forge official seals/letterheads.
Research vs drafting
- Research — Suggest search terms, help locate public document titles/agencies, and cluster themes — Humans retain responsibility for the factual basis of published work
- Drafting — Outline, draft, edit for clarity — Own the publish decision; ensure labels (fact/analysis/speculation; deployed/prototype/speculative)
- Fact claims — May propose wording — Humans retain responsibility for the wording used in publication
- Images — May generate illustrative art when disclosed — No fake “declassified photo” or false documentary imagery
The default public claim is: drafting may be AI-assisted; humans retain publish responsibility. Do not claim a formal newsroom multi-editor review.
Disclosure placement (EU transparency–aware)
Minimum disclosure stack (recommended):
- Sitewide AI Use Policy at /ai-policy/ (this page).
- Author page AI blurb (see Adriano Margarone author materials).
- Article-level short notice on AI-assisted pieces (trust block or byline area)—not footer-only.
- Footer disclaimer may continue for continuity but must not be the only notice.
Image AI: if an image is AI-generated, label it in caption or adjacent text (e.g. “Illustrative image; AI-generated”). Do not present AI art as archival government photography.
Images
- Prefer real documents, official stills, and properly licensed photography.
- AI-generated images are for illustration only, must be labeled, and must not fake seals, letterheads, crash debris “photos,” or documentary authenticity.
- Avoid EXIF/GPS surprises on user uploads (see Privacy Policy).
Accountability
- Publisher / site control: Insider Release (Adriano Margarone, owner/founder).
- Author-attributed pieces: Named human author bears public responsibility for the byline; AI remains a tool.
- Corrections: If AI contribution led to an error, correct under the Corrections Policy; do not blame the model in place of fixing the record.
Questions about this policy and error reports may be sent to editor@insiderrelease.com as an interim contact.
Relationship to EU expectations
Insider Release aims to treat AI-assisted content with transparency, clear labeling, and human editorial responsibility. Exact legal classification of each feature or article type can depend on jurisdiction and use case; this page is editorial policy, not legal advice.
Disclosure wording
This article may include AI-assisted drafting. Editorial responsibility for publication remains with Insider Release or the named editor.