How we generate credibility scores

First, chrome extension users highlight suspicious information in their news articles, sending it along to Public Editor.

An expert annotator receives and reads the article, highlighting problematic passages mapped to categories of misinformation. Passages are sent to specialist annotators for further review.

Specialists in each category assess the problematic passages in close detail, identifying specific mistakes.

Work of the specialists is statistically evaluated. When multiple specialists find the same mistake type, that mistake is factored into the article’s credibility score.