Bengaluru techie fires cook after AI monitoring system catches her stealing fruit

A domestic cook in Bengaluru lost her job after an AI-powered home surveillance system used by her employer detected her taking fruit from his house, provoking ethical consternation about the use of AI monitoring tools in private homes. What happened Pankaj Tanwar, a Bengaluru-based tech professional, developed and deployed an AI monitoring system in his kitchen suspecting theft by his cook, who earned around ₹4,800 monthly. A CCTV camera integrated with vision and Anthropic's Claude Haiku language model, Tanwar's so-called "AI roommate" provided granular logs of his cook's movements, tracking her arrival time, monitoring how frequently she washed her hands, and evaluating the thoroughness of her cleaning (e.g., noting that the area behind the stove had been neglected). The system recorded her stealing 3 apples, a banana, and some blueberries, and failing to wash her hands properly. Tanwar fired her after receiving a weekly summary and shared the details publicly on X, igniting a fiery backlash over disproportionate punishment and surveillance. Why it happened The incident was driven primarily b y a drastic power imbalance between tech-savvy employers and domestic workers with few formal rights and little bargaining power in India . While Tanwar stated he had confronted the cook twice previously without result, leading him to "innovate" a solution , he deployed the system covertly and without meaningful consent. It was further compounded by the broad accessibility and low cost of powerful and invasive AI technologies and tools. What it means For Tanwar's cook, the incident resulted in the immediate loss of her job, livelihood, and dignity. For society, it signals a shift towards normalising a "panopticon" in the home, with domestic workers subjected to 24/7 forensic scrutiny, often opaquely and with little or no recourse. For policymakers, the incident highlights a regulatory vacuum: while labour laws in India cover formal workplaces, the domestic sphere remains a legal and ethical "grey zone", indicating a need for guidelines on informed consent, proportionality (is firing someone for stealing an apple a proportionate response?), and data privacy for domestic staff. System 🤖 AI roommate Developer: Pankaj Tanwar Country: India Sector: Domestic ser vice Purpose: Domestic surveillance, inventory tracking, hygiene monitoring Technology: Computer vision; Generative A I Issue: Accountability; Consent; Power inbalance; Privacy/surveillance; Propo rtionality; Transparency Timeline ⏰ February 2026: Tanwar installs monitoring system. Late February 2026: The system monitors the cook for two weeks; she initially behaves cautiously but eventually ignores the camera. March 1, 2026: Tanwar receives alerts and a weekly report from his "AI roommate" documenting the fruit being taken, and hygiene lapses. March 2, 2026: Tanwar announces on X that he has fired his cook.

Date it happened

2026-03-01

Organisation involved

Pankaj Tanwar

Product, system or model

AI roommate

This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.