Last updated: 25 October 2026
Where the data comes from
Every event on this site originates in the GDELT Project’s public event and knowledge graph feeds, which index news coverage worldwide and update every fifteen minutes. We ingest those feeds continuously and store the events we can attribute to a reachable source article.
We do not republish articles. Each story page carries a headline, extracted context, and a link to the original publisher, who retains all rights to their reporting.
How stories are selected
Selection is automated and based on measurable properties of the coverage, not on editorial preference:
- Volume relative to baseline. An event is promoted when coverage exceeds what is normal for its country and category over the preceding period.
- Tone magnitude. Strongly positive or strongly negative coverage is surfaced ahead of neutral reporting.
- Source breadth. Events reported by several independent outlets rank above those carried by one.
- Recency. Older events fall out of the live feed regardless of their other properties.
We do not accept payment to feature, promote, suppress or remove a story, and we have no commercial relationship with any outlet whose reporting appears here.
Use of AI, and how it is labelled
Summaries on high-signal story pages are generated by a language model running on hardware we control. The process is:
- We identify stories that pass the signal threshold — a small fraction of daily volume.
- We retrieve the full text of the source articles, from several outlets where more than one has covered the event.
- The model is given only that retrieved text and asked to summarise what happened and why it matters. It is not asked to recall facts from memory, speculate, or fill gaps.
- The result is stored and displayed under a heading that identifies it as an AI summary.
AI output is always labelled. We never present machine-written text as the work of a human author, and we never attach a byline to it.
AI is never the only source. The original article link appears on every story page, above and independent of any summary.
Automated summarisation makes mistakes. A model can misread a quotation, confuse two similarly named people, or compress a qualified statement into an unqualified one. If you find an error, the correction process below applies in full — we treat an AI error exactly as seriously as any other.
Corrections
Write to us via the contact page with the story URL and what is wrong. We aim to respond within three working days.
- Factual errors in our summary are corrected, or the summary is withdrawn.
- Errors in the underlying reporting are the publisher’s to correct. Tell us anyway — if a source article is retracted, we remove the corresponding page.
- Automatically generated headlines that misrepresent a story are corrected, and where the fault is systematic we fix the rule that produced it.
Removal requests
If you are named in a story and want the page removed, write to us. We will remove pages where the underlying article has been retracted or removed, where a legal right to erasure applies, or where the page causes identifiable harm that outweighs its public interest. We will not remove accurate coverage of matters of genuine public concern simply because it is unwelcome.
Independence and funding
World Echo News is independently operated and not affiliated with any publisher, government or political organisation. Where advertising appears, it is clearly distinguishable from editorial content and has no influence over what is published or how it ranks.