Adolescents’ Interpretations of Romantic Chatbots, Relational Control, and AI-Enabled Harm

Abstract

This exploratory qualitative study examines how adolescents interpret AI-mediated intimate practices that may not yet be widespread among minors but are already salient within their digital imaginaries, focusing on romantic chatbots, AI-assisted relational advice, and AI-enabled practices of control and harm. Six online, gender-segregated focus groups were conducted with 33 adolescents aged 16–18 in Spain. Participants framed romantic chatbots as sites of idealized, conflict-free intimacy shaped by user control; evaluations of AI-based relational advice were ambivalent, particularly among girls, who questioned its personalization and intent. Adolescents condemned highly intrusive practices such as message cloning, yet often normalized other forms of monitoring (e.g. real-time location sharing) as safety or care. Deepfakes were consistently identified as the most harmful practice, with participants highlighting violations of consent and noting gendered dynamics in peer-group circulation. Overall, marked AI appears less as a source of novel inequalities than as an amplifier of existing affective and gendered scripts in adolescent digital cultures.