"Machine Learning Can Help to Flag Risky Messages on Instagram While Preserving Users' Privacy"
It is important to protect social media users from harassment and bullying while also taking steps to protect their privacy. A team of researchers from four leading universities has proposed using Machine Learning (ML) technology to identify potentially risky conversations on Instagram without eavesdropping on them. The discovery could provide platforms and parents with the ability to protect vulnerable, younger users while maintaining their privacy. The team led by researchers from Drexel University, Boston University, Georgia Institute of Technology, and Vanderbilt University recently published their work on investigating what type of data input, such as metadata, text, and image features, could be most useful for ML models to identify risky conversations. Their findings suggest that risky conversations can be identified based on metadata characteristics, such as conversation length and participant engagement. Afsaneh Razi, Ph.D., an assistant professor in Drexel's College of Computing and Informatics and co-author of the study, stated that the prevalence of harassment, abuse, and bullying by malicious users is very concerning, considering Instagram's popularity among young people. Instagram makes its users feel safe enough to connect with others very openly. After the Cambridge Analytica scandal and the European Union's precedent-setting privacy protection regulations, platforms are under increasing pressure to protect their users' privacy. Therefore, Meta, the company behind Facebook and Instagram, is implementing end-to-end encryption for all platform messages, indicating that the message content is technologically protected and can only be accessed by those involved in the conversation. However, this increased degree of security makes it more difficult for platforms to use automated technology to detect and prevent online threats, which is why the team's system could play a crucial role in protecting users. This article continues to discuss the system developed to use ML to help flag risky messages on Instagram while preserving users' privacy.