Skip to main content

Trust and transparency

Editorial and corrections policy

This policy explains how public TrendsAGI research is produced, what our measurements mean, and how to report a correction.

Last materially reviewed:

Publication standards

Evidence before publication

Automated trend briefings must be tied to a stored signal snapshot, a finite measured posts-per-hour rate, and at least two independently identified public sources. If those checks fail, the system does not publish the draft.

Clear limits

A snapshot describes what our collection pipeline observed at a stated time. It is not a forecast, a count of unique people, or a guarantee that the same rate is still current. Articles distinguish measured values from interpretation and state when a cause is uncertain.

Automation disclosure

Automated briefings are labelled on the page. They are generated from stored measurements and source-backed research and are not represented as manually reviewed. File-backed editorial guides are labelled separately.

Search quality controls

Dynamic pages that do not meet our source and measurement requirements remain readable but are excluded from discovery feeds and marked noindex. We update an existing topic where practical instead of creating keyword variants with substantially the same content.

Sources and attribution

External factual claims should link to the pages used to support them. We prefer primary records, official documentation, original research, and direct statements. Two pages from the same publisher do not count as two independent sources. A link shows what informed the briefing; it does not imply that the publisher endorses TrendsAGI.

Updates and corrections

We may update a briefing when a material fact, source, or interpretation changes. Meaningful changes update the page's modification date. We do not change dates merely to make older work appear fresh.

To report a factual error, send the page URL, the statement at issue, and supporting evidence through our contact page. Confirmed material errors are corrected as promptly as practical; a correction note is added when context would otherwise be misleading.

Measurement methodology

Our research page documents the detection pipeline, velocity and acceleration concepts, threshold calibration, source diversity, and known limitations.

Read the research methodology