Fabula AI “able to identify 93 percent of ‘fake news’ within hours of dissemination” - TechCrunch
TechCrunch released an in-depth analysis of Fabula AI’s fake news detection product on 6 Feb 2019, quoting the best-in-class accuracy of 93% (ROC AUC).
Fabula’s ground-breaking “truth-risk scoring platform”, being trained on months of data from the likes of Snopes and Politifact, is based on the diffusion characteristics of the news in a social network - making it both language and locale independent. This 93% accuracy is achieved within a few hours of the news first appearing, this is critical to the next phase of Fabula’s journey, when the platform will be opened up (using APIs) to platforms and publishers who can integrate the technology.
Read more of TechCrunch’s analysis, see: https://techcrunch.com/2019/02/06/fabula-ai-is-using-social-spread-to-spot-fake-news/