How we use AI
AI is at the core of how we research and draft. Atlas, our research agent, gathers material and produces first drafts grounded in primary sources. AI is the engine, not the final authority: a named human editor is accountable for what we publish. Where a page was assembled by template-substituting facts from a structured source, we disclose that on the page itself, so you always know how a given piece was produced.
Sourcing and citation
Every substantive claim is tied to a source you can check — UN documents, treaty text, foreign-ministry statements, official statistics, and verified analysis. We prefer primary sources over secondary summaries and link to the original wherever possible. We do not publish authority we cannot trace: if a claim cannot be sourced, it does not run.
Human oversight and authorship
Pages that make substantive claims carry a named author or editor, listed in our author directory. People — not models — decide what we stand behind, review claims against their sources, and own the result.
Independence
Our research follows the evidence. Coverage is not shaped by advertisers, sponsors, or any government or institution. When we describe contested positions, we attribute them to the party that holds them rather than presenting them as settled fact.
Corrections
When we get something wrong, we fix it in the open and record what changed. See our corrections policy for what we correct and how to report an error.