BBC News built Juicer to monitor approximately 130,000 social media posts daily across multiple platforms. The system flags potential breaking news by identifying unusual spikes in geographic clusters or keyword combinations that suggest developing stories.
How the Filtering Process Works
Juicer scores posts based on account credibility, location data, and correlation with other reports from the same area. When a building fire starts in Manchester, the system detects multiple posts from verified local accounts within a 500-meter radius within minutes. It ranks these above single unverified claims.
The Human Verification Layer
Every story flagged by Juicer goes to BBC verification journalists who contact posters, reverse-image search photos, and confirm details through official sources. During the 2017 Manchester Arena attack, Juicer identified early eyewitness posts within four minutes, but BBC editors spent 23 minutes verifying facts before publishing. The AI shortened discovery time from hours to minutes; it did not eliminate verification requirements.
Student Interpretation
Beginners assume AI can fact-check and verify information automatically, making verification specialists obsolete.
Professional Reality
Verification experts recognize AI handles volume screening impossible for humans but cannot assess source credibility or confirm facts. The BBC still employs the same number of verification journalists as before Juicer. The tool changed what they spend time on: less manual social media scrolling, more direct source contact and fact confirmation. The system amplified human capability rather than automated judgment.