Fabula AI Detects Fake News by Analyzing Social Spread Patterns
UK startup Fabula AI introduces a new method for identifying fake news that bypasses the need to understand content and instead analyzes how information spreads across social networks.
UK startup Fabula AI introduces a novel approach to combating disinformation on social media by avoiding the complex task of parsing text altogether. Instead of trying to teach artificial intelligence to read and understand news content, the company focuses entirely on how information travels across social networks.
The startup utilizes a patented form of geometric deep learning to study complex, non-Euclidean datasets like social graphs. This technology identifies distinct differences in the spreading patterns of false information compared to genuine news, looking at the trajectory and the users involved in the distribution.
By monitoring these characteristic propagation patterns, Fabula AI claims it can accurately detect both malicious disinformation and unintentional misinformation. This method offers a potential workaround to the massive technical challenges that major platforms like Facebook face when trying to moderate billions of posts using traditional content analysis.