DiscussionJoin free
PokerAPT Taipei Day 1A draws 412 entries, Day 1B starts at noonPokerLive: 188 of 412 remain, Pete Chen leadsSportsSpurs vs Lakers tonight 08:30TechA freelancer books 14 AI jobs in a monthEntertainmentThe Poker Apartment episode 1 is liveGamingChain game withdrawals tested: real numbers from 12 playersPokerToday’s perk: free entry to the APA Wednesday freerollPokerAPT Taipei Day 1A draws 412 entries, Day 1B starts at noonPokerLive: 188 of 412 remain, Pete Chen leadsSportsSpurs vs Lakers tonight 08:30TechA freelancer books 14 AI jobs in a monthEntertainmentThe Poker Apartment episode 1 is liveGamingChain game withdrawals tested: real numbers from 12 playersPokerToday’s perk: free entry to the APA Wednesday freeroll
Back home

AI Guidelines

How we use AI ourselves, and what we never do with it.

Last updated: 10 September 2026

This site has a Tech section that writes about AI, so readers have a right to know how we use it ourselves. This page describes what we actually do, not what we aspire to.

1. Judgement and words are a person’s responsibility

Our reporting is researched and written by an editor. AI tools may be used to organise public information, find sources, translate, catch typos and suggest headlines. **Deciding what is worth reporting, whether a fact holds up, and how to write it is the editor’s job and is not delegated to a model.**

2. When we label

  • Content whose main passages are AI-generated is labelled on the article.
  • AI-generated images are labelled as such and are never mixed with real photography without saying so.
  • Translation: parts of the English, Japanese and Korean editions are machine-translated. **Where versions differ, the Traditional Chinese version prevails** — particularly for the Terms, the Privacy Policy and this page.

3. What we never do

  • Use AI to fabricate events, quotes or scenes that did not happen — not even labelled as AI-generated.
  • Use AI to generate a real person’s face, voice or movements (deepfakes), including in entertainment and gaming coverage.
  • Feed readers’ personal data, unpublished interview material or partners’ confidential information into external AI services.
  • Use AI to manufacture reader comments, reviews or votes.

4. Model output is not a source

Generative models produce plausible-sounding errors and can reproduce other people’s work without attribution. So we treat AI output as a lead, never as a source: every fact goes back to the original, and every image’s licence is checked.

5. The site’s code

Part of this website’s code was written with AI assistance. That does not affect editorial content, but since this page is about transparency, it belongs here.

6. This will change

The technology and the industry’s standards are both moving. We will update the revision date when this page changes. If you think we have fallen short, write and tell us — the point of publishing this is that it can be checked.