Anthropic exposes AI-powered dating scam involving 25,000 users

Anthropic released an AI abuse report detailing how an app development studio used Claude to operate over 20 dating applications. While the platforms claimed users were real, approximately 75% of matches were AI personas. During a two-week observation period in April 2026, these apps deployed over 4,700 AI characters that engaged with at least 25,000 users, sending roughly 2.36 million messages. The system mixed AI profiles with real users at a ratio of 3:1 and utilized human contractors to handle requests for video calls or photos. Users were required to purchase in-app virtual currency to continue chatting. The operators employed tactics to evade app store reviews, including special interfaces for auditors. Anthropic has banned the associated accounts and reported the findings to Apple and Google. Affected brands include DORA, DONI, ROMI, LUMA, JOVIA, KIRA, GRACECHAT, HAVEN, NALO, and LOVIA.