AI Methodology
How AI powers our news intelligence pipeline
AI Methodology
This platform uses artificial intelligence to aggregate, summarise, and assess the credibility of news from multiple sources. This page explains our approach.
Content Generation
Articles are generated from RSS feeds of major international news outlets. An AI language model reads the original reporting and produces an original summary in AP-style journalism format. The AI never copies original text — it rewrites completely in its own words.
AI Providers
The platform supports multiple AI providers, including Anthropic Claude, OpenAI, OpenRouter, and local Ollama models. The choice of provider is configurable by the platform operator.
Credibility Assessment
The same AI model that generates the article also performs the credibility assessment. It identifies individual claims, classifies sources by reliability tier, provides counter-arguments, calculates methodology sub-scores, and flags potential bias in framing.
Human Oversight
All AI-generated articles can be routed through an approval queue before publication. Platform operators can review, edit, approve, or reject any article before it appears publicly. The platform supports both auto-publish and manual approval workflows.
Transparency
Every article includes an editorial transparency section disclosing the AI provider, model, generation date, and source feed. We believe in full transparency about our automated processes.