This is a good point. Instead of policing what trolls will use it for, the same AI should be able to detect racist content and prevent it from spreading