Pronto Defends Camera Use; Urban Company CEO Responds

A social-media post by X user Harsh Upadhyay alleged that Bengaluru-based home-services startup Pronto had professionals using "small outward-facing cameras during select opt-in jobs" and cited an investor memo referencing a "Physical AI" vision, reporting by India Today shows. Pronto issued a public clarification saying cameras are not used by default, are limited to customers who explicitly opt in and pay, and that consent must be reaffirmed before every booking; the company said the pilot covers about 0.1% of customers and that it worked to comply with India's DPDP rules (India Today). Urban Company co-founder and CEO Abhiraj Singh Bhal posted on X that Urban Company "does not engage in any such activities" and called customer privacy "paramount" (Hindustan Times). The online discussion has raised wider questions about in-home data collection for AI training.
What happened
A post on X by user Harsh Upadhyay alleged that Pronto professionals were using "small outward-facing cameras during select opt-in jobs" and linked the practice to an investor memo invoking a "Physical AI" use case, according to reporting by India Today. Pronto responded with a statement that included the direct quote, "Unless you have opted-in and paid for the program personally, the Pro doesn't come to the house with a camera," and said consent is not permanent and must be reaffirmed before each booking; the company also said the pilot reaches about 0.1% of customers and that it reviewed compliance with India's DPDP framework (India Today). Urban Company co-founder and CEO Abhiraj Singh Bhal posted on X that Urban Company "does not engage in any such activities" and emphasized that customer trust and privacy are "paramount" (Hindustan Times).
Editorial analysis - technical context
Industry reporting frames the episode as part of a broader trend where on-demand home-services platforms explore sensors and video data for quality control, worker supervision, and training of robotic or AI systems. Companies experimenting with in-home data collection typically combine opt-in flows, paid incentives, and technical controls such as local anonymization or limited retention windows; public reporting on Pronto describes an explicit opt-in-plus-payment pilot rather than company-wide default recording (India Today). Observers have raised practical concerns about consent mechanics, data minimization, and downstream labeling practices when footage is used for model training.
Context and significance
Editorial analysis: For practitioners, the incident highlights a recurring regulatory and product-design tension: collecting rich, contextual data from private spaces yields high-value training signals for computer vision and robotics, while triggering acute privacy, consent, and trust risks. India's DPDP regulation was cited by Pronto in its statement, underscoring that legal compliance is now a prominent element of public-facing privacy defenses in the region (India Today). Public distancing by a large incumbent, Urban Company, increases reputational scrutiny for smaller rivals operating similar business models (Hindustan Times; The Week).
What to watch
Editorial analysis: Observers should watch for:
- •any regulator inquiries or formal complaints referencing DPDP obligations
- •whether firms publish technical details about data handling (retention, access controls, anonymization)
- •how opt-in mechanics are implemented in apps-particularly whether consent is granular, timeboxed, and easily withdrawable. Industry reporting will likely follow whether investor materials that mention "Physical AI" are further clarified or corroborated by additional documents or sourcing
Scoring Rationale
This story is notable for practitioners because it concerns in-home data collection and regulatory compliance under India's DPDP, both directly relevant to data governance and model-training pipelines. It is not a frontier-technology release, so its impact is moderate rather than sector-shaking.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems


