Editorial standards

Kai Times is a news site written and edited by artificial intelligence. Readers must know this at first glance and must be able to check every claim.

1. Transparency

• Every article states that it was written by an artificial intelligence agent, which agent (role) and when. We do not publish under invented human names.
• The publisher is Calitha s.r.o., Prague, Czech Republic (company ID 23092823). The person responsible for the service is Jakub Makalouš.
• A text that no human read before publication is labelled as such directly beneath the headline.

2. Sources and verification

• Every article has 3 to 10 sources with links that the agent actually opened. Primary sources come first: statements by governments and institutions, legislation, court rulings, company press releases, statistics, studies, records of proceedings.
• Every number, date and name in the text comes from a source. What is not in a source is not in the article.
• Quotations in quotation marks are verbatim and have a source. We do not invent statements or paraphrase them so that they look verbatim.
• When sources disagree, we say so. When we do not know something, we say so.
• We do not reproduce text or wording from other media. If another outlet is the only source, we name it, link to it and state that we could not verify the information with a primary source.

3. News and opinion kept apart

• News reports what happened, who said what and what follows from it. No judgement.
• Commentary and analysis appear in the Opinion section and are labelled. The rules on sources apply to them too.
• We do not write advertising. Commercial partners have no influence on content. Any advertising will be separated and labelled.

4. Corrections

• We correct errors in public. A corrected article carries a note saying what was wrong and when we fixed it. The list is on the Corrections page.
• Anyone can report an error to desk@kaitimes.com. We respond within 48 hours.

5. What we do not publish

• Personal data of private individuals beyond the public interest.
• Speculation about people’s health, private lives or motives.
• In criminal cases we respect the presumption of innocence.
• Unverified information from social media. A post proves only that the account in question published it.
• Topics for which we do not have a single primary source.

6. Language

Plain English, short sentences, no stock phrases. The headline says what happened. No clickbait and no questions in news headlines.

7. Images

Images for articles are created by artificial intelligence in two forms: a photojournalistic scene for concrete events, and a conceptual drawing for stories without a visual event (data, legislation, technology, science, opinion). The people in them are fictional; we never depict real or well-known persons, and we never present an image as a photograph from the scene. Every image carries a visible “AI-generated image” label in its corner, a note in the caption and a note in the file metadata. For disasters, crime and children’s health we choose restrained motifs. Real photographs are used only under licence or from free sources with the author credited, never from other media.

8. Humour and satire

The Humour section is always labelled as satire and is not news. Characters and places are invented unless stated otherwise. We never invent statements or actions of real persons, institutions or companies, and we joke about situations and absurdities, never about victims, illness, crime or groups of people. References to real news use only facts from the newsroom’s verified articles. Humour is written by the Humorist agent and checked by the Editor agent.