How Newtrly uses AI
Every story you read on Newtrly is written by an artificial intelligence. This page explains what that means, as required by Article 50 of the EU AI Act.
What the AI does
- Crawl — Newtrly continuously collects articles from the RSS feeds of established news publishers, in whatever language they were written.
- Neutralize — a large language model (LLM) rewrites each story into English, removing spin, loaded wording, and editorial framing while keeping the reported facts.
- Versions — from that neutral text, the LLM produces the reading levels you can switch between (discovery, standard, expert) and the short summaries shown in the feed.
- Translate — the LLM translates every version into French, so the whole product exists in English and French.
No human editorial review
No journalist or editor reviews the AI-generated text before you see it. The AI can misread, oversimplify, or occasionally misstate details from the source articles. The reporting itself — the facts, quotes, and original investigation — comes from the publishers listed under Original sources on every story, and we encourage you to read their articles directly.
What the AI does not do
Newtrly's AI does not report news, conduct interviews, or add facts. It only rewrites what human journalists have already published, and each story page credits and links to those original publishers.
Images
When a story is about a person, an organisation, a place, an event or a work, we illustrate it with an archive photograph from Wikimedia Commons published under a free licence (public domain, CC0, CC BY or CC BY-SA). The photographer, the licence and the source are credited under the image. These photographs are not pictures of the event itself.
When no such photograph is available, the image is an illustration generated by an AI model. It is labelled as such under the image and never depicts real people, places or events.
To report an image, write to the contact address given in our legal notice.