Description
Full Fact AI is a suite of tools built by Full Fact, a UK-registered charity (no. 1158683) and non-profit company (no. 06975984) best known as an independent fact-checking organisation. Full Fact AI is the technology side of that work: rather than checking claims one story at a time, it's built to monitor public debate at scale and flag the statements worth a fact-checker's attention.
Read More
The platform's Automatic Claim Detection reads and listens across everything a team monitors, and its Claims view ranks statements by a SIFT score and relevance, updating in real time. Users can filter by custom categories, focus on claims not yet reviewed, or step back through past dates to see how a narrative developed. Relevant claims then get a full SIFT analysis: an AI-generated accuracy verdict, a harm score out of 100, and a breakdown of how misleading, believable, actionable, and consequential the claim is, alongside source-linked corrections and key facts.
Teams tell Full Fact AI what to watch by adding keywords and topics, and the tool scans RSS feeds, newspapers, YouTube, TikTok, Instagram, podcasts, other social media accounts, and broadcast radio for matches. A Slack integration sends a notification, with the matched claim, its source, and a review link, whenever something matches a tracked topic; claims can be starred, dismissed, or muted so a team focuses on what matters.
Full Fact AI is used by fact-checkers, journalists, researchers, and communicators. Full Fact states it is trusted by over 45 organisations in 30 countries, including Africa Check, Dubawa, and The Stage Media Liberia (vendor-reported figures). The platform currently supports English, French, and Arabic, with more languages in development.
Beyond the core product, Full Fact also runs related projects: Prebunking at Scale, built with the European Fact-Checking Standards Network and Maldita.es to catch misleading short-form video narratives across Europe before they go viral; MediaVault, a searchable library of fact checks originally built by Duke Reporters' Lab and now hosted by Full Fact; and an AI Harms Assessment Tool built with AISI. There is no public pricing; access is by requesting a demo through the website.
- CategoryAI Research Assistants
- Pricing ModelFree
- DeploymentCloud
- PlatformsWeb
Full Fact AI's Core Features
Automatic Claim Detection that reads and listens across monitored sources to surface checkable statements
Claims view ranking statements by SIFT score and relevance, updating in real time
SIFT analysis per claim: an AI-generated accuracy verdict plus a harm score out of 100
Source monitoring across RSS feeds, newspapers, YouTube, TikTok, Instagram, podcasts, and broadcast radio
Custom keyword and topic configuration to focus monitoring on specific subjects
Slack integration that alerts teams when a tracked claim appears, with source and review link
Filtering, starring, and muting of claims to manage alert volume
Date-based history view to track how a claim or narrative evolved over time
How to use Full Fact AI?
Request access: Contact Full Fact AI through the website to request a demo (there's no public self-signup)
Set up sources: Add the keywords and topics you want monitored across news, social media, and broadcast
Review the Claims view: Check statements ranked by SIFT score and filter by category or date
Read the SIFT analysis: Open a claim to see its harm score, accuracy verdict, and supporting facts
Connect Slack: Link Slack to get notified when a new matching claim appears
Act on it: Star, dismiss, or mute claims as your team reviews and responds
Full Fact AI's Use Cases
- Newsroom fact-checking
- Political speech monitoring
- Cross-border misinformation tracking
- Fact-check research
