Olympics’ online trolls get taken down by AI

Artificial intelligence-based technology that tracks abusive sentiment in social media messages, including texts, images and even emojis, has prevented 4000 individually targeted threats towards Olympics competitors in the first week of the games, with details of some senders passed on to police around the world for potential prosecutions.

The Threat Matrix technology, built by London-based data science company Signify and deployed by the International Olympic Committee, has previously been used at other major sporting events – it led to a Queensland man being arrested and charged for threatening a referee and his family on social media during last year’s Rugby World Cup.