Automated Multimodal Detection and Reporting of Cyber Bullying using ML and Encryption

Cyberbullying refers to the harassment, threatening or damage using online platforms. There was a significant increase in the number of cyberbullying cases mainly due to a spike in the number of people using the internet. According to AAG IT Support, there is a victim of cybercrime every 37s, making it an average of 97 victims per hour; however, only few of them are actually reported to the cyber cell. Despite the alarming issue, many cases go unreported due to several reasons. Hence, there is a need for an automated cyberbullying reporting system to be employed so as to increase awareness and help the victims. With accuracy rates of 98.5%, 88.7%, and 80.7%, respectively, this study uses machine learning techniques like Random Forest for text-based identification and ensemble models for identifying cyberbullying in photos and videos. To guarantee data security and integrity, cryptographic techniques including RSA for safe key exchange, SHA-256 for tamper-proofing, and AES for encryption are used. The current reporting methods are not properly streamlined, resulting in delayed responses for victims and hence there is crucial need for an automated system. Thus the approach proposed aims in identifying any kind of cyber bullying and automatically reports it, creating a safer digital environment for all the users.