AI Actress Calls Face-Scan Callers and Track Mood

The recent viral moment featuring AI actress Tilly Norwood glitching into Chinese during a television interview has drawn global attention to her creators’ video calling service. Behind the novelty of speaking with a synthetic character lies a significant shift in how digital interactions are policed. To access the "Talking Tilly" platform, users must now pass a mandatory age verification check that involves scanning their face, a requirement that applies to callers regardless of their location.

Once the call begins, the system does not merely listen to the conversation. It actively analyzes the user's visual cues and tone of voice to infer emotional states, allowing the AI to adjust its responses accordingly. This level of monitoring is not optional and cannot be disabled for individual calls, marking a distinct trade-off for those seeking this specific type of interactive entertainment. The service is set to shut down permanently in late September, creating a limited window for this particular blend of technology and surveillance.

Mandatory Scans for Age Verification

Before a single word is spoken, callers must submit a video selfie to an identity verification provider. This process estimates the user's age to ensure compliance with adult content regulations. If the automated estimate is unclear, the system requires a government-issued photo ID. The company states that no biometric templates are stored, and images are deleted after the check, retaining only an approximate age band and a reference number. This gate was added recently to the terms of service, aligning the platform with stricter UK online safety standards that have begun affecting adult websites and social media platforms.

Continuous Monitoring of User Emotions

During the call, the platform processes live audio and video feeds to detect mood shifts. This data is used to tailor the AI’s dialogue, ensuring the character responds in a way that matches the caller's perceived emotional state. The privacy policy explicitly states that this feature cannot be switched off. Legal basis for this processing relies on "legitimate interests" rather than explicit consent, a choice documented in the company’s version history. While this creates a more engaging experience, it means the user’s facial expressions and vocal patterns are continuously analyzed without a direct opt-out mechanism.

Service Limitations and Data Retention

The service operates on a time-based model, offering five free minutes before requiring payment for additional time. However, all purchased minutes expire when the service shuts down on September 27, meaning any unused time is forfeited. Transcripts of calls are retained for up to eight weeks and may be reviewed by staff or third-party partners. Additionally, the AI retains memory of previous conversations to personalize future interactions, though this data can be deleted upon request. As reported by BleepingComputer, these terms highlight a growing trend where regulatory compliance drives the adoption of invasive privacy measures in consumer tech.