Last updated: 18 August 2026.
This page is the single, canonical statement of how Wehpa uses artificial intelligence. It exists so you never have to guess whether a human was involved in something you are reading or looking at. It sits alongside our Editorial Standards and Corrections pages, and it is the version that governs if the three ever disagree.
The short version
- A named human is accountable for everything published here.
- AI never supplies a fact, a quote, a number or a date.
- Illustrations made with AI are labelled as illustrations, on the image itself.
- We do not generate images of real, identifiable people.
- We do not use your personal information to train AI models.
How we use AI in writing
We use AI as a drafting tool — the way a newsroom uses a style guide or a spellchecker, not the way it uses a reporter. It does not decide what we cover, and it is never the source of a fact.
Specifically, AI is never the origin of a quote, a statistic, a date, a chart position, an award, or any claim about a living person. Those come from a named source, checked by a human. Every proper noun, number and date in a draft is treated as unverified until someone has put it against a source.
Articles drafted with AI assistance carry a disclosure line at the foot of the piece. The named writer remains responsible for the article regardless of what tool helped produce a sentence.
How we use AI in images
Wehpa covers history, food, music and culture. Much of what we write about was never photographed, and a great deal of what was photographed sits behind restrictive licences. Rather than illustrate a nineteenth-century dockside or a pre-Columbian griddle with a stock photo that shows something else entirely, we commission an illustration and generate it.
Every AI-generated image on this site carries a visible caption:
An artist’s interpretation by Wehpa, created with AI — not the real thing.
Its alternative text, which screen readers announce, begins with the word Illustration. The label travels with the image, so it survives being shared, embedded or scraped.
What we will not generate
- Real, identifiable people. No living public figure, no historical figure, no private individual. Where a person is the subject of an article, the image carries the object, the place or the document instead.
- Anything presented as documentary. We do not generate images that could be mistaken for a photograph of a real event, and we do not use AI images to depict news.
- Fabricated evidence. No invented documents, records, headlines, screenshots or archival material.
- Legible text. Generated images that produce readable words are regenerated rather than retouched, because generated text invents things.
Where the images come from and who owns them
Our illustrations are generated from prompts we write ourselves, and we keep those prompts on file. We do not upload third-party copyrighted images to be transformed. Where an image on the site is a real photograph rather than an illustration, it is either licensed, public domain, or used with permission, and it carries an attribution caption instead of the AI disclosure line.
Your information and AI
We do not use your personal information — your email address, your comments, your account, your reading history — to train AI models, and we do not sell or supply it to anyone else for that purpose. See our Privacy Policy for what we do collect and why.
Regulatory position
We label AI-generated visual content voluntarily, and have done so since we started using it. That labelling is also intended to satisfy the transparency duties for AI-generated and manipulated content set out in Article 50 of the EU Artificial Intelligence Act, which began applying on 2 August 2026, and comparable obligations elsewhere. We treat labelling as an editorial commitment first and a compliance obligation second — if the two ever diverge, we will keep the label.
When we get it wrong
If you find an image on Wehpa that is unlabelled, mislabelled, appears to depict a real person, or reads as a photograph when it is not, tell us: info@wehpa.com. We will correct it and record the correction on our Corrections page rather than fixing it quietly.
We began applying this standard in July 2026. We are working back through the older archive to caption and, where necessary, remove images that predate it. That work is ongoing and incomplete, and we would rather say so than imply the archive is already clean.
Changes
This is a fast-moving area, both technically and legally. We will update this page as our practice changes and update the date at the top when we do.