Upgrade Your Artificial Intelligence
Essential Training Data
The quality of artificial intelligence relies upon robust, versatile training data. For over 160 years, Reuters has built a legacy as a trusted source of accurate, bias-free information, and the quality of our news data reflects this.
With sophisticated metadata found in over 45 million pieces of fully-licensable news content, Reuters material is an indispensable source of training data, whatever your project is.
Be the first to know
Want to stay up to date on our machine learning & AI updates? Subscribe to our email updates in two easy steps and you’ll never miss a beat.
Why is Reuters News Data Different?
Rights: Reuters has the proprietary rights to our data corpus and visual assets
Trust & Accuracy: Over 2000 media companies rely on Reuters news to make editorial and business decisions every day. Guided by Reuters Trust principles, our news preserves integrity, independence and freedom from bias
Diversity: Broad coverage of major topics from over 200 global locations and 16 languages, including business, finance, politics, sports, entertainment, technology, and much more
Metadata: Our advanced metadata contains regional and category-specific codes, allowing for intelligent grouping
New Data Use Cases
30 million articles, 16 languages. Our extensive body of newswires is readily translated into multiple languages with parallel sentencing.
Advanced image metadata promotes the detection of action and events within our collection of over 13 million images.
Speech to Text
Training Closed-captions and video transcripts in over 800K video assets facilitate advanced training to identify figures, scenes and soundbites.
Entity-rich data gives contextual value and promotes the building of sophisticated knowledge graphs.