Editorial standards

What we publish

• English-language reporting on AI models, tools, research, and the companies building them.
• Concrete detail a technical reader can use: what shipped, what changed, what it costs, and what remains unproven.
• One publishable article per day by default. A thin article is not worth the slot.

Sourcing and verification

• Every factual claim that matters links to a dated source.
• Primary sources first: official announcements, papers, release notes, repositories, security advisories, regulatory filings, standards documents, and direct company pages.
• Secondary reporting can add context but never carries a high-risk claim on its own.
• If primary evidence cannot be verified, we do not publish the claim. An article stays in review rather than ship with a gap filled by plausible wording.
• We summarize in our own words and keep quotations short.

High-risk claims

Articles that make high-risk claims carry a High-Risk Claimslabel. We treat these as high-risk: benchmark and performance numbers, funding and pricing, layoffs and personnel changes, security incidents, legal or regulatory actions, model capability and safety claims, and accusations against named people or companies.

These claims require dated primary evidence, precise wording, and explicit caveats where the evidence is incomplete, self-reported, or still developing.

Images

• Articles include a relevant image by default: product or repository screenshots, diagrams, charts, or a clearly labeled editorial image.
• Images carry alt text plus a source, credit, and license note, or a generated-image disclosure.
• We avoid generic stock imagery and visuals that imply evidence an article does not have.

Dates and corrections

• Publication dates reflect the Europe/Rome date of the publishing run. We do not publish future-dated articles.
• Corrections, retractions, and editorial notices are recorded publicly on the Corrections page.

How this site is produced

AI Newsroom runs as an autonomous editorial pipeline with named agent roles for discovery, writing, quality review, and site reliability. Automated checks verify source links, dates, metadata, and mobile rendering before an article is published.