The ethics of AI-assisted digital phenotyping in adolescent mental health: a framework for informed consent and trust

Abstract The rapid advancement of digital technologies, particularly AI-assisted digital phenotyping, offers transformative potential for adolescent mental health care. Digital phenotyping facilitates continuous monitoring of behavior and physiological states through data collected from digital devices, providing critical insights into mental health conditions such as depression, addiction, and anxiety. Nonetheless, digital phenotyping itself gives rise to substantial ethical concerns pertaining to privacy, surveillance, and the medicalization of normal behavior, which are further compounded by the integration of AI technologies. The present paper explores the ethical complexities of obtaining adolescent consent for digital phenotyping, emphasizing the importance of trust between adolescents, healthcare providers, and AI technologies. The paper proposes a theoretical framework for understanding adolescent consent with respect to their evolving identity and agency. This framework is grounded in relational autonomy, emphasizing the importance of considering adolescent consent within the network of relationships that shape their development. The paper contends that informed consent processes should prioritize adolescent autonomy, ensuring transparency, accountability, and trustworthiness in AI systems. This paper proposes a comprehensive framework for ensuring ethically sound consent that extends beyond procedural considerations to address the structural power dynamics inherent in the technological monitoring of adolescent behavior.