The expansion of artificial intelligence within social media platforms has generated new forms of cyber delinquency in Iran, ranging from transformed methods of offending to the erosion of legal and social control mechanisms. This study aims to identify the role of artificial intelligence in the cyber delinquency of Iranian social media users. Employing a meta-synthesis methodology and drawing on documentary analysis of scholarly sources, official reports, and empirical evidence, the research synthesizes theoretical perspectives from media and sociological frameworks, including social control, symbolic interactionism, surveillance capitalism, filter bubbles, and social learning theory. The extracted themes were classified into nine major domains: privacy violations, phishing and digital extortion, dissemination of fake news, production of deceptive content (deepfakes), facilitation and tutoring of online criminal acts, creation of fake accounts, hate speech, media addiction, and algorithmic discrimination. The findings demonstrate that these phenomena can be analyzed within three overarching spheres: threats to individual security, disruption of social order, and legal-cultural challenges. Ultimately, the comprehensive concept of “the weakening of trust structures and social control in the digital sphere” is proposed as the central conclusion. The results indicate that in the absence of effective regulatory policies and public education, artificial intelligence can become a tool for facilitating delinquency and undermining social cohesion. This study provides a foundation for rethinking AI governance, expanding media literacy, and informing cultural and legal policymaking in virtual environments.