AI use policy
What the machine does, and what a person is responsible for.
We use AI tools in producing this site, and we'd rather be specific about how than hide behind a vague “AI-assisted” label.
What AI is used for
- Gathering and organising candidate sources for a place, relationship or historical event.
- Drafting the plain-language explanations from those sources.
- Structuring data (timelines, relationship diagrams) and routine copy editing.
What AI is not allowed to do
- It does not decide what is true. Every factual claim — dates, split-vs-overlay, which code came from which — is checked by a named editor against the cited source before it is published.
- It does not invent sources, quotations, boundaries, or “value” scores.
- It does not publish. Drafts do not go live automatically; a person reviews and approves.
- It is not given an opinion. This site reports cited facts; it does not run machine-generated commentary presented as a view.
Where responsibility sits
A real, named editor is accountable for verification, editorial review and corrections — see the Editor page. “The site checked it” means a person checked it, not that a model was asked to be careful.
If we ever find that AI-assisted content slipped through with an unverified claim, we correct it openly under our Corrections policy and treat it as a process failure to fix, not to paper over.