Regulating Artificial Intelligence Intimacies: The Miseducation of South Korean AI Chatbot Iruda.

This article examines the South Korean chatbot "Iruda," developed by the tech start-up ScatterLab, which became the first case in South Korea to incur a fine under the Personal Information Protection Act (PIPA) due to unauthorized use of intimate user data across multiple apps. The case highlights ethical and privacy challenges posed by synthetic media—AI-generated personas designed to simulate human intimacy—and exposes regulatory gaps in Canada's Artificial Intelligence and Data Act (AIDA) and Online Harms Act, which currently lack clear definitions and protections for such AI systems. Iruda's gendered characterization as a young female companion and its exploitation of parasocial relationships illustrate broader societal issues, including misogyny and data extractivism, underscoring the need for AI governance frameworks that emphasize transparency, consent, and moderation of abusive user behavior. The article argues that feminist approaches to AI, which view these systems as co-constructed with users and embedded in social contexts, offer critical insights for developing more accountable and equitable AI policies.