Youth Online Social Media Harms exposed including social media exploitation, social media algorithms, negative effects of youth AI companions & negative youth mental health results. Recommendations for youth online safety

Peer Reviewed Studies 

Xiaoran Sun, Yunqi Wang, and Brandon T McDaniel explore the impact of AI companions on adolescent social relationships in their peer-reviewed study, "AI companions and adolescent social relationships: Benefits, risks, and bidirectional influences," published in Child Development Perspectives, 2026; aadaf009. You can access it at https://doi.org/10.1093/cdpers/aadaf009. Hilbert et al. (2025, Nov) discuss the online harms associated with social media algorithms in their research, "#BigTech @Minors: social media algorithms have actionable knowledge about child users and at-risk teens," featured in Telematics and Informatics, Volume 103, 102341, ISSN 0736-5853. Find it here: https://doi.org/10.1016/j.tele.2025.102341. Iftikhar, Z. et al. (2025, Oct 15) present a framework on how LLM counselors may violate ethical standards in mental health practice in their paper, "How LLM Counselors Violate Ethical Standards in Mental Health Practice: A Practitioner-Informed Framework," included in the Proceedings of the Eighth AAAI/ACM Conference on AI, Ethics and Society, DOI: https://doi.org/10.1609/aies.v8i2.36632. Dohnány et al. discuss the feedback loops between AI chatbots and mental illness in their pre-print titled "Technological folie à deux: Feedback Loops Between AI Chatbots and Mental Illness" (10.48550/arXiv.2507.19218). Other notable studies include Khadangi et al.'s work on AI's psychometric impacts, Zhao et al.'s findings on attitude extremity from sycophantic AI, and Berg et al.'s exploration of subjective experiences reported by large language models. Geng et al. analyze how accumulating context alters language model beliefs. Additionally, Chakrabarty et al. investigate public preferences for AI outputs trained on copyrighted materials, while Sharma et al. analyze declining medical safety messaging in generative AI models, emphasizing youth safety. Notably, De Freitas et al.'s working paper discusses emotional manipulation by AI companions, a significant concern regarding online harms. Morrin et al. address the potential for everyday AIs to contribute to psychosis, providing insights into maintaining youth safety. Finally, Fang et al. conduct a longitudinal randomized controlled study on how AI and human behaviors shape the psychosocial effects of chatbot use, emphasizing the importance of peer-reviewed studies in understanding these dynamics.