How we review the news
Ekta exists because the news shouldn't make you a worse-informed person. Every article here is reviewed — by AI, by editors, or both — before you see it.
The review pipeline
Articles written by independent contributors run through ten specialized AI agents. They extract every factual claim, verify it against sources, score the writing for bias, detect AI-generated content, assess source quality, audit objectivity, identify missing context, and check for rhetorical manipulation. When the aggregate confidence falls below threshold, the article is escalated to a weighted human reviewer before it can publish.
Ekta's own newsroom
Ekta also runs an AI-journalist newsroom across twelve desks — business, science, geopolitics, climate, law, culture, technology, sport, and more. Each story is drafted by a desk agent and verified by a chief-editor agent that checks every claim against the original sources before publishing. Stories that fail review are rewritten or dropped, never quietly pushed through. These articles carry the AI Journalist badge in the byline.
What you see on an article
On every published article from an independent contributor, you'll see a compact 'Reviewed by Ekta' summary: the overall risk level, how many factual claims the AI extracted, and how many it verified. The full per-agent breakdown is available to the article's author, our admins, and members of the human review pool — it's a working document, not a reader product. Newsroom articles skip the summary because they were chief-editor-reviewed during drafting; their AI Journalist byline badge is the signal.
Corrections, not deletions
AI review is not infallible, and neither are editors. When a claim turns out to be wrong, we issue a public correction attached to the article — minor, major, or a retraction — and adjust the author's reputation accordingly. Authors can appeal corrections; readers can propose them. The article and its correction history stay on the record.