AIooze is researched and drafted by AI agents, and every piece is reviewed by an editor before it goes live. We think a site about AI should be open about how it uses AI, so this page sets out what the agents do, what the editor does, and where the limits are.
The agents
Each agent has one job, runs on a schedule and is built on commercially available AI models.
- News scout (daily). Sweeps for what happened in AI in the last day or two and shortlists the strongest stories.
- News drafts (daily). Checks each story against its primary source and drafts three to five short briefs, tagged by topic and country, each linking to the original.
- Timeline check (weekly). Reads the week’s published news and proposes the few stories, if any, that belong on the AI Timeline. Most weeks it proposes none.
- Price check (weekly). Compares every model and tool in the Directory against the vendor’s own pricing pages and reports anything that has changed.
- Explainer review (monthly). Rereads each Learn AI explainer against the month’s news and flags anything that has gone out of date.
- Timeline health check (daily). Checks the timeline is current and complete, and flags entries waiting for review.
What the editor does
Nothing publishes automatically. The agents file drafts and reports; an editor reads them and decides what goes live, what goes back for another pass and what gets dropped. Directory prices and explainer updates are changed only after that review.
The rules the agents follow
- Primary sources only: vendor announcements, official documentation, regulator filings and research papers. Aggregators, listicles and content farms are excluded, because this field is full of AI-generated sites that invent model names and figures.
- No invented facts, figures, quotes or dates. If sources disagree, the piece says so.
- Every piece links to its source, so you can check it yourself.
- Short, plain and unhyped. Briefs are written fresh and never copy text from the sources they cover.
- Stories are tagged where they happened, so the country hubs reflect the whole world, not just Silicon Valley.
The limits
AI agents make mistakes. They can misread a figure, compress a claim too far or miss context a specialist would catch. The review step and the source links are there to catch that, but some errors will get through. If you spot one, email hello@aiooze.com and we’ll correct the page. Our corrections policy explains how.
Why we work this way
AI is good at reading widely and consistently, every day. That’s most of what keeping up with this field takes. We use it for that, keep a review step before anything is published, and show you how the site is made rather than pretend otherwise.