AI actress hotline demands video selfie for age verification before any call
A viral AI actress named Tilly Norwood, who appeared on a talk show and glitched into Chinese, offers a video-call service that requires callers to submit a video selfie for automated age estimation before the first connection. The service also analyzes callers' facial expressions and voice tone during calls to gauge mood, and records and transcribes conversations using US-based providers. The company behind the service says it retains only an approximate age band and reference number, not the selfie or biometric data, but the mood analysis cannot be disabled for individual calls.
The age verification relies on Didit's facial estimation, with a government ID as a backup if the scan is unclear. Xicoia claims no biometric template is stored, only an approximate age band and a reference number. The mood analysis, which runs on every call, uses camera and audio cues, and its legal basis is "legitimate interests" rather than explicit consent, a change made this month.
The service is a limited engagement, ending September 27, with all purchased minutes expiring then. A safety classifier incorrectly flagged a benign weather conversation as abusive, though recordings are deleted after 24 hours. Transcripts are kept for up to eight weeks, and the AI retains conversation memory for personalization, which users can request to delete.
This service could accelerate the normalization of biometric age checks for mainstream AI interactions, extending a practice currently limited to adult sites. It may also set a precedent for emotional surveillance, as the always-on mood analysis could become a standard feature in companion chatbots, raising privacy concerns. The false-positive content flagging suggests automated moderation may overreach, potentially chilling free expression in casual conversations. Younger users, in particular, may face new barriers to access, while adults might accept these trade-offs for novelty.