Data Security on Social Media Under Threat for Training AI Modules
Atharva R. Khot,
Mansi D. Agrawal
Student,
Department of Information Technology S. D. S. M. College, Palghar, Maharashtra, India
Abstract :
Intelligence (AI) systems increasingly rely on large-scale data drawn from social media platforms. While these datasets enable powerful predictive models, they also introduce substantial risks to privacy, security, and user consent. This research paper explores the vulnerabilities associated with social media data when used to train AI modules. Drawing on existing scholarship, we highlight the ethical, legal, and technical challenges posed by unauthorized scraping, re-identification, and governance gaps. Through the works of Robertson et al., Frichot, and Gerbrandt & Howard, we emphasize how current practices undermine user trust and call for stronger data governance and privacy-preserving mechanisms. The paper concludes that without meaningful safeguards, AI development risks exacerbating digital exploitation and eroding privacy rights.
Keywords: Data Security, Social Media, AI Training, Privacy, Governance, Ethics
