A computer-implemented method including scanning, by one or more processors, data that is maintained on one or more social networks, identifying, one or more social entities associated with the scanned data, determining a risk score for each of the one or more social entities, analyzing the scanned data using one or more machine learning techniques, assigning a rating to each of the one or more social entities based on the determined risk score and the analyzed data, and determining, based on the rating assigned to at least one of the one or more social entities that the at least one social entity is a security risk.