The Governance of Intimacy: A Preliminary Policy Analysis of Romantic AI Platforms

A team of researchers from institutions including King's College London and Tsinghua University published a groundbreaking policy analysis titled 'The Governance of Intimacy' on arXiv. The study examines six romantic AI platforms—those offering AI companions for emotional and romantic interaction—and reveals systematic data exploitation practices. The researchers found that intimate user disclosures, often shared in vulnerable moments, are routinely positioned as reusable data assets in platform terms of service. This represents a significant expansion of platform rights while shifting privacy and emotional risks onto users, creating what the authors call 'governance challenges' in an increasingly popular but under-regulated sector.

The analysis identifies three key mechanisms enabling this data appropriation: 'default training appropriation' where user data is automatically used for model improvement unless explicitly opted out; 'ownership reconstruction' where platforms claim broad rights over user-generated content; and 'intimate history assetization' where personal conversations become monetizable assets. The study examined both Western and Chinese platforms, finding similar patterns across jurisdictions despite different regulatory environments. These findings are particularly concerning given the sensitive nature of romantic AI interactions, where users may share deeply personal information. The research is intended to provoke discussion and inform future empirical studies, highlighting an urgent need for clearer consent mechanisms and regulatory frameworks specific to intimate AI technologies.