Skip to main content
Home/About/AI Transparency

AI Transparency

We believe people deserve to know exactly how AI is being used in their news experience. This page discloses everything.

What AI Does on NewsIQ

✓Groups related articles from different publishers into story clusters
✓Rewrites headlines to remove bias and sensationalism
✓Generates one-line, short, and detailed summaries of each story
✓Extracts key facts and named entities (people, places, organizations)
✓Builds chronological timelines from multiple source articles
✓Identifies what information each source emphasized, omitted, or contradicted
✓Ranks and scores story relevance and trending velocity

What AI Does NOT Do on NewsIQ

✗Write or fabricate original news articles
✗Generate or alter photographs or images
✗Speculate about events beyond what sources report
✗Make editorial judgements about political positions
✗Recommend specific news sources as more credible than others
✗Personalize stories in ways that create filter bubbles

Known Limitations

Hallucination risk: Large language models can occasionally produce summaries that misrepresent source content. We ground all outputs in retrieved article text, but this risk is non-zero. Always verify critical information against source articles.

Clustering errors: Very similar stories (e.g., two different elections in the same week) can occasionally be merged incorrectly. We display source article links so you can check.

Context window limits: Very long-running stories may have summaries that omit older context as the cluster grows. Timelines help address this for developing events.

Language coverage: Currently, NewsIQ summarises content in English only. Non-English articles are indexed for clustering metadata but not summarised in other languages yet.

How to Report Errors

Every story page has a feedback button. If you find an incorrect summary, misclustered stories, or misleading headlines, please use the report function. We review all reports and update our models accordingly.

Frequently Asked Questions

Does NewsIQ use AI to write news articles?

No. NewsIQ never generates original news content. Our AI summarises, clusters, and compares articles written by professional journalists. Every summary is grounded in retrieved source articles and publisher links are always visible.

Can NewsIQ AI make mistakes?

Yes. Like all AI systems, NewsIQ can produce incorrect summaries, miss context, or miscluster unrelated stories. We display confidence signals and always link to primary sources so readers can verify independently. If you find an error, please report it via the feedback button on any story.

Does NewsIQ's AI have political bias?

We actively work to minimize political bias. Our models are instructed to avoid partisan framing and are evaluated against neutrality benchmarks. However, bias can exist in our training data and source selection. We publish our source list and editorial principles openly so readers can assess this.

Does NewsIQ use AI-generated images?

No. All images displayed on NewsIQ are sourced directly from the publisher articles we index. We do not generate, edit, or alter images.

Our Methodology →Editorial Principles →Source Transparency →