AI, Dating, and Romantic Relationships

Gen Z college students are leaving dating apps in record numbers while AI features are being added to those same apps at unprecedented rates. We ask under what conditions AI integrations in dating apps appeal to Gen Z college students whose disengagement is driven by swipe fatigue, perceived inauthenticity, and the desire for organic connection. The study uses a mixed-methods design built around a single question: what people said they wanted from AI in dating, and whether they would still make the same choice when the scenarios pushed them in different directions. Twenty pre-vignette semi-structured interviews surfaced what participants said they wanted; a vignette experiment with forty-five participants then tested those positions by presenting twelve swipe decisions across three AI integrations (AI-curated matching, AI conversation coaching, AI-assembled group meetups), each followed by a three-step counter-pressure ladder that added friction for those who initially accepted or reassurance for those who initially rejected; twenty-nine participants supplied open-ended debrief reflections. The interview content was organised under five themes mapped one to one onto the study’s framework: three exit factors (swipe fatigue, perceived inauthenticity, desire for organic connection), AI platform usage, and AI concerns. Baseline acceptance rates lined up with what participants said in the interviews (matching 80%, group meetups 67%, conversation coaching 49%), but most participants did not stick with their initial choice when pushed: reassurance moved more rejecters into adoption (73 to 87 percent flip rates) than friction broke acceptors out (47 to 82 percent), and the directions in which participants flipped (auto-reply imitation, opaque algorithms, surveillance) align with the themes that emerged in the interviews. The paper translates these findings into five design interventions grounded in interpersonal-privacy and privacy-by-design literature for apps (Petronio, 2002; Nissenbaum, 2010; Hartzog, 2018; Marwick & boyd, 2014; Hancock, Naaman, & Levy, 2020) and in established usability and transparency standards (Nielsen, 1994; ISO 9241-110:2020; IEEE 7001-2021), and discusses theoretical, design, and policy implications. The contribution is an empirical account of which AI integrations in dating apps Gen Z students accept, which they reject, and the conditions that distinguish the two.