Current social media monitoring only highlights ‘False Negatives,’ which produces a large number of results. However, the percentage of the true value of these results is low. This is because existing forms of detection avoid labeling text as not interesting in the context of toxicity when, in actual fact, it is, as the sinister intention is masked upon initial inspection due to the context surrounding the message being excluded, implying the message has a positive nature. An additional factor of subversion is the disguising of hate speech using positive language. The ability to detect positive toxicity through its use of natural language processing (NLP) makes the WhoDis Project unique, which is why NLP is so important. Coupled with background contextual information, NLP allows for the detection of the context and structure of the language used, helping us find the origins of hate campaigns and conspiracy theories.
Following the Justice for Prosperity project, we aim to uncover subversion, expose silent polarisation, and use intelligence and security for good, untangling the web that has blinded those who have been trying to protect our society’s most vulnerable from hate.