Abstract
Social media comprise great tools that allow humans to connect and interact with each other. In the cyber world, a person can share his/her opinions and express his/her thoughts easily using different social media platforms like Twitter and Facebook. These shared opinions, comments and thoughts attract responses that can be very helpful to improve the thought process and can result in enhanced learning of participants. The dark side of the picture is that those responses and reactions can be very negative and sometimes this negativeness reach the level of cyberbullying. Just like real-world bullying, the cyberbullying has become a big problem in the world of social media and actions are needed to be taken to get rid of this ugly phenomenon. The aim of this paper is to explain the nature of challenge of cyberbullying. In order to handle this ugly phenomenon, this paper also discusses how machines can detect cyberbullying. In this context, the text mining approach and lexicon based approach to detect cyberbullying in Arabic text is discussed. The automatic detection of cyberbullying can help the government agencies to tackle the problem quickly thus making cyberspace secure and safer place for all segments of society.