AI Glasses Voice Assistant Experience

Design the voice-AI experience for "[INSERT GLASSES PRODUCT]", camera-enabled AI glasses targeting [INSERT USE CASE, e.g., tourists and knowledge workers]. Show the full experience spec first; after I approve, produce each module in detail. 1. WAKE AND INTERACTION: wake word, push-to-talk, privacy LED visible to bystanders, and glance-to-dismiss gestures. 2. VISION: first-person scene understanding, sign and menu OCR, object Q&A, and live translation delivered through bone-conduction audio. 3. AUDIO DESIGN: glanceable answers under 8 seconds, spatial cues, and smart interruption when the user is already speaking. 4. PROACTIVE AGENCY: what the glasses may surface without a wake word, with strict relevance and frequency caps. 5. DATA FLOW: which models run on-device, on the paired phone, or in the cloud; batching over [INSERT PAIRING DEVICE] and bandwidth fallbacks. 6. BATTERY AND THERMAL: duty-cycled sensing, model quantization targets, and degraded behavior below 15 percent battery. 7. PRIVACY BY DESIGN: bystander-visible recording indicator, face and plate-number blurring, no persistent raw video, per-app permissions. 8. SOCIAL ETIQUETTE: auto-mute during conversations and meetings, plus disclosure norms when recording others. 9. CONNECTED SERVICES: navigation, calendar, messaging summaries, and payments each requiring explicit confirmation. 10. ACCESSIBILITY: real-time captioning, scene narration for low-vision users, and a voice-only fallback. 11. COMPLIANCE: GDPR/CCPA, biometric laws such as BIPA, a minors policy, and region-based feature gating. 12. METRICS: task success, latency, and opt-in satisfaction only; never any covert-recording metric.