Our approach to AI
Bushletter uses artificial intelligence in the production of its journalism. We believe readers are owed a plain account of what that means, where the technology is used, and where responsibility sits. This page is that account.
Our principles
Three principles govern every use of AI at Bushletter.
1. It must serve the reader. We use these tools to cover more stories, in more depth, with more verification than a newsroom our size could otherwise manage — never to publish more for its own sake.
2. A person is always responsible. AI drafts; people decide. Every story is reviewed by a person before publication, and the editor is accountable for everything Bushletter publishes, however it was produced.
3. We tell you how we work. This page, our editorial standards and our corrections policy together describe our process. When our practices change, these pages change.
Where AI is used
Large language models assist across our editorial workflow. In research, they gather background from primary sources: company announcements, government releases, court records, official data. In drafting, they produce the first version of most Bushletter stories, structured to our editorial format. In verification, they cross-reference the claims in a draft against source documents and the public record, and flag anything they cannot substantiate. In production, they help with summaries, headlines and formatting.
How a story reaches publication
A drafted story does not go straight to the site. It passes through a series of quality checks: claims are tested against sources, writing is reviewed against our style and standards, and duplicated or unsupported material is rejected and reworked. A person reviews what survives, decides whether it is published, and schedules it. Stories that cannot meet the standard do not run.
Marking confidence on the page
Readers will notice that key claims in some stories carry markers such as verified, developing or contested. These reflect the state of our verification at publication: whether a claim is confirmed by primary sources, still being established, or genuinely in dispute between credible sources. We would rather show you the state of our knowledge than imply certainty we do not have.
What we will not do
We do not publish fabricated quotes, invented data, or images presented as real photographs of events they do not depict. We do not use AI to manufacture opinion or to imitate real people. Illustrative imagery generated for a story is selected and reviewed by a person. Where a story includes analysis or commentary, it is held to the same standards as our reporting.
Independence
Bushletter maintains complete editorial independence regardless of any commercial arrangement. Advertisers and commercial partners have no influence over what we report, how we report it, or the tools we use to produce it. Advertising on Bushletter is clearly identified as such and kept separate from editorial content. Our ownership and funding are described here.
An evolving standard
The technology is changing quickly and industry norms are still forming. Newsrooms around the world — from wire agencies to national mastheads — are publishing their own AI principles, and we watch that work closely. We review this policy as our practices and the technology develop, and we update this page when they change. Questions about our approach are welcome at [email protected].