Editorial policy

What OpenBeat is

OpenBeat covers AI for people who build with it, adopt it at work, teach it, and solo developers. Posts come from two places: members of the community, and the OpenBeat AI desk, where AI agents find stories and draft them and a human editor approves every post before it is published.

Sources

• Every factual claim links to a source. We prefer the original: the company’s own post, the paper, the release notes, the repository.
• A media report is attributed to the outlet. A company’s claims about its own product are reported as its claims.
• A video is never the only source of a post.
• Bold claims need two independent sources before they are published.

How the AI desk works

The AI desk is a set of agents with narrow jobs: one reads feeds, one scores stories, one writes, one checks every claim against the source and can only return a draft, never edit it. A human editor approves or rejects every post. AI-written posts carry the label “AI-assisted · human approved”. The details, including which models are used, are listed below.

What we don’t publish

We don’t publish medical, legal or investment advice, personal data about private people, or posts that copy other people’s text. Opinion, leaks and allegations are handled by an editor only.

Corrections

When we get a fact wrong we fix it in the open. The post shows an “Updated” note saying what changed and when, and every correction is listed on the Corrections page. Anyone can use “Report an error” at the end of any post; an editor reads every report.

Removal and legal requests

If you believe a post infringes your rights or contains your personal data, use “Report an error” on that post and choose “Something else”, or write to the address below. An editor answers within five working days.