AI Disclosure

Our policy on AI-assisted journalism.

What we do with AI

• Draft news summaries from curated public sources (press releases, regulatory filings, conference announcements, vendor blogs).
• Suggest headlines and SEO metadata for human editors to refine.
• Cluster related news for our weekly digest.

What we don't do with AI

• We don't publish AI output without human review. Every AI-drafted article is read, fact-checked, and edited before publication.
• We don't fabricate quotes, sources, or details. If an AI hallucinates, it's our editors' job to catch it — and if we miss one, we correct and disclose.
• We don't use AI to reproduce copyrighted material. Source text is paraphrased; citations link back to originals.
• We don't use reader data to train AI. Reader comments, accounts, and email addresses never leave our infrastructure for AI training purposes.

How to spot AI-assisted articles

Every AI-assisted article carries an “AI-assisted” badge next to the byline and lists its source URLs at the end of the article. Articles written entirely by a human carry no such badge.

Which AI we use

We use Anthropic's Claude model via the Anthropic API, with prompts engineered to require citations, refuse paywalled or low-quality sources, and paraphrase aggressively. No reader personally identifying information is ever sent to Anthropic.

Corrections

Spot something wrong? Email corrections@retailmedia360.com. We correct errors promptly and note substantive corrections at the top of the article.