Policy

AI Transparency Policy

Version 1.0 · effective October 12, 2026

We use AI as a drafting assistant, not an oracle. Here is exactly what it does and does not do.

Where AI is used

  • Proposing topics and drafting outlines and articles from retrieved sources.
  • Cross-checking draft claims against those sources.
  • Suggesting titles, descriptions and internal links.
  • Generating the featured photograph for each article (no real people, no text).

Where it is not used

  • AI does not invent sources, statistics, experts or personal stories; our checks reject drafts that do.
  • AI does not write first-person experience. Bylines credit the editorial team, not fictional authors.
  • No reader data is sent to AI systems.

Models and providers

Research, drafting and fact-checking use Anthropic's Claude (Claude Code on the operator's own subscription); images use OpenAI image generation through the operator's Codex/ChatGPT plan. Each article records which model and prompt version produced it.

Human oversight

Publishing is automatic. Articles are not individually reviewed by a human before they go live; instead every article must pass automated source, fact and safety checks, and anything that fails is held back. Readers who spot a problem can report it and we will correct it.

Disclosure

AI-assisted articles carry a "How this article was made" note. If you spot an error, tell us at [email protected].