An Employer Deployed Covert AI Surveillance at Home, Then Fired His Cook Over Three Apples
What happened
In March 2026, a domestic cook in Bengaluru lost her job after her employer, Pankaj Tanwar, received a weekly summary from an AI-powered home surveillance system he had installed two weeks earlier without telling her. The system logged her taking three apples, a banana, and some blueberries from the refrigerator, and flagged that she had not washed her hands properly. Tanwar fired her and shared the full details publicly on X, where the post triggered a backlash over both the scale of the surveillance and the proportionality of the response.
The system, which Tanwar called his "AI roommate," used computer vision and generative AI to produce detailed logs of the cook's behavior throughout her working day. It tracked her arrival time, how often she washed her hands, and how thoroughly she cleaned specific areas of the flat, noting, for example, that the area behind the stove had been neglected. The cook was reportedly cautious at first after noticing the camera, but gradually resumed her normal routine. She did not know the system was generating weekly forensic reports on her movements.
The incident drew immediate criticism not because AI surveillance is unusual in commercial settings, but because it happened inside a private home, where the cook had no formal right to be told she was being monitored, no union to file a grievance with, and no access to the data collected about her. The power asymmetry runs entirely in one direction: a technically capable employer who configured, interpreted, and published the system's output, facing an informal worker who could neither contest nor examine it. Tanwar stated he had spoken to the cook twice before installing the system, but the deployment was covert and unconsented regardless of what preceded it.
India's labor laws govern formal workplaces, but domestic employment sits in a gap those laws do not reach. There is no requirement for informed consent before deploying surveillance inside a home where workers are present, no proportionality standard governing what level of conduct justifies what level of consequence, and no data protection mechanism giving a domestic worker the right to know what has been collected about her. The incident was entirely legal, which is much of what makes it worth examining.
The accountability gap the case exposes is not primarily about whether the cook stole fruit or whether Tanwar had grounds to terminate her. It is about the absence of any verifiable record of what was agreed to before the camera went up, what standards governed what the system flagged, and what review occurred before a weekly AI summary resulted in a permanent job loss. A provable record of what a system did, under what terms it was deployed, and what a worker was told before it started watching her would not have saved the job. But it would have forced the question of consent to the surface before the harm, rather than after.
Reported impact
- Affected parties
- Not publicly disclosed
- Harm type
- Not publicly disclosed
- Scale
- Not publicly disclosed
- Financial impact
- Not publicly disclosed
- Regulatory action
- Not publicly disclosed
Classification
Relevant governance controls
Governance control mapping is not available for this record.
- No controls mapped
Not publicly disclosed
Control mapping is analytical. It does not state that any control would have prevented the incident.
Sources and evidence
This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.