AI Content Policy
Last updated
We write about AI, so you are entitled to know exactly how much of this site it wrote. Here is the whole answer.
The short version
A human decides what to write, tests everything that claims to work, and edits every word that reaches you. AI is used the way a spellchecker and a research assistant are used — as tools inside that process, never as a replacement for it.
Where we do use it
- Research and outlining. Gathering what has already been said on a topic, so we can find what has not.
- First drafts of routine sections. A definition, a step list, a summary of something uncontroversial.
- Editing. Tightening sentences, catching repetition, flagging jargon that slipped through.
- The prompts themselves. Obviously. They are written, run, broken, and rewritten until they work.
Where we do not
- Anything presented as tested. If a page says a prompt works, a person ran it. There is no exception to this and there will not be one.
- Anything presented as experience. “We tried this for a month” means somebody did.
- Facts and figures. Every number is checked against its source by a human, because this is the thing these tools are worst at and most confident about.
- Reader quotes and reviews. Never generated, never composited, never “representative”. If we cannot get permission to publish a real one, the section stays empty.
- Bulk publishing. We are not producing a hundred articles a month. The whole point of this site is that somebody read it first.
Why we tell you
Because the alternative is a site that writes about AI while being cagey about its own use of it, and that is not a site worth trusting on the subject.
Also because it is a real difference. A great deal of what currently ranks for these searches is generated, unread, and untested, and you can usually tell about two paragraphs in. If you ever cannot tell here, something has gone wrong and we would like to know.