Artificial intelligence (AI) companion chatbots are rapidly becoming part of teens' daily lives during critical stages of identity formation and development. Many of the AI companion chatbots are highly customizable and interactive. There is evidence that these chatbots can expose users to developing emotional attachment, excessive use, and overreliance. Current safety measures for AI chatbots are often inadequate or easily bypassed. This CAREER project seeks to address these issues by designing, developing, and assessing developmentally appropriate interventions to foster healthy AI interactions among teens. It will benefit society broadly by addressing the mental health of teens as impacted by emotional and cognitive attachment and overreliance on AI companions. It will enhance digital literacy and critical thinking among teens and their caregivers, helping them engage with AI technologies in more effective ways. The interdisciplinary research plan seeks to transform knowledge at the intersection of digital safety for teens, developmental psychology, AI ethics, and human-AI interaction through three main objectives. The first objective is to develop a novel empirical foundation that bridges teens' self-reported data and behaviors through triangulated research methods, examining their experiences with AI companion risks and parasocial relationships, as well as the ways they cope and build mental resilience against the potential negative impacts of AI. The second objective is to create reusable teen-centric sociotechnical design frameworks and a suite of interventions to enhance AI companion safety through co-design activities involving teens and their safety stakeholders. The third objective is to evaluate and longitudinally assess the age-appropriate suite of novel interventions in real-world contexts. The project builds upon well-established theoretical approaches to develop datasets, models, and tools. The project's activities include plans to benefit the pu