Cameras, Consent & AI: What Should Be Allowed?

Discussion Session AI systems are increasingly learning from the physical world - from workers sewing clothes and building walls to professionals performing tasks inside our homes. Companies are collecting first-person video, or “egocentric data”, to train robots and Physical AI systems. At the same time, services like Pronto Verified are bringing cameras into private homes, raising new questions around consent, privacy, surveillance, ownership, and compensation. This discussion asks a simple but difficult question: What should be allowed when human activity becomes training data for AI? We’ll explore:• What does meaningful consent look like when participation happens at work or inside the home?• Who owns the data generated by a person’s movements, skills, and labour?• Is anonymization enough to protect privacy?• Should workers be compensated for the long-term value of the data they generate?• Can people realistically refuse when cameras are introduced by an employer or service provider?• What safeguards should exist as Physical AI and robotics scale?• Where should we draw the line between useful AI data collection and surveillance? Join us for an open discussion on AI, labour, privacy, consent, and the emerging economics of human-generated data. Reading Links (Must read before joining the session) The Guardian: Indian factory workers told to film themselves for AI2. Al Jazeera: India's workers are training AI robots to take their jobs3. Economic Times: Pronto's premium 'Verified' service with wearable camerasJoin the WhatsApp community for quick updates.Organised By AI Safety India