The Effect of Social Media Engagement Optimization Algorithms on Minor Users' Mental and Cognitive Health in Correlation to Academic Performance: Adaptive Feedback Loop Detection and Mitigation in Social Media - A Machine Learning Approach

Social media platforms employ sophisticated engagement optimization algorithms designed to maximize user attention and profit. This sociotechnical thesis addresses the research question: How do social media engagement optimization algorithms contribute to prolonged screen time in minors, and how does this affect their attention span and mental health in relation to school performance?

My technical project proposes a feedback loop mitigation system with three core components: a Pattern Recognition Module that analyzes engagement patterns, an Intervention Engine that implements strategies including timed breaks and content diversification, and a User Dashboard providing transparent visualization and control. The system aims to detect potentially harmful algorithmic feedback loops and implement targeted interventions while maintaining user agency and privacy.

Using a mixed-method approach combining meta-review analysis of academic literature with discourse analysis of user experiences across platforms like YouTube, TikTok, and Reddit, my STS research examines how these algorithms operate at the intersection of computer science, behavioral psychology, and corporate profit motives to impact adolescent cognitive development.

Engagement optimization algorithms deliberately target developmental vulnerabilities in the adolescent brain through variable-ratio reinforcement scheduling that creates persistent dopamine-driven engagement cycles. These neurological mechanisms directly conflict with academic requirements by training the brain to prioritize brief, unpredictable rewards over sustained attention tasks. Studies reveal algorithm-driven content significantly impairs working memory function among younger users whose prefrontal cortex development remains in earlier stages. These algorithms simultaneously create contradictory behavioral patterns in adolescents through what researchers refer to as "antisocial design." While platforms publicly promote connection, their operational profit mechanisms rely on emotionally provocative content that generates higher engagement metrics. Meta-analyses identify significant correlations between extensive social media usage and lower academic performance, documenting disruptions to study patterns that undermine learning capacity.

Beyond individual psychological mechanisms, social media algorithms embed themselves within adolescent social spheres, transforming platform usage from optional entertainment into perceived social necessity. Platforms create heightened expectations around friendship maintenance where teenagers feel obligated to publicly validate peers' online activities, fundamentally altering relationship dynamics. These social obligations become particularly problematic when algorithms amplify harmful content, as adolescents must choose between psychological well-being and social inclusion.

My technical and STS projects work in tandem to address these complex challenges - developing both technological solutions to mitigate harmful feedback loops while advancing our understanding of how these systems fundamentally reshape adolescent cognitive development and academic performance. This integrated approach recognizes that addressing algorithmic harms requires both technical intervention and deeper societal understanding of how these systems operate.

School of Engineering and Applied Science

Bachelor of Science in Computer Science

Technical Advisor: Briana Morrison

STS Advisor: Karina Rider