Bengaluru Startup Pronto Faces Backlash Over Home‑Video AI Training

When you invite a tech company into your living room, you probably expect a quick demo, not a hidden camera rolling while you pour coffee. Yet that’s exactly what happened to several users of Pronto, a Bengaluru startup that markets an AI‑powered home‑assistant service.

Earlier this week, a handful of customers posted screenshots of video clips that had been recorded inside their apartments without explicit consent. The footage, they say, was meant to "train" Pronto’s machine‑learning models—essentially teaching the AI how people move, speak and interact with everyday objects.

The revelation sparked an immediate outcry on social media. Users accused Pronto of breaching privacy norms, and some even hinted at potential legal action. “I felt watched,” one homeowner wrote, “and it’s not the kind of surveillance I signed up for.”

Pronto quickly responded, publishing a statement that tried to calm the storm. The company claimed the recordings were part of a “voluntary pilot program” and that participants had allegedly given consent when they signed up for the beta. It added that the videos were stored securely, anonymised wherever possible, and would be deleted once the AI models were sufficiently trained.

Critics, however, aren’t buying the excuse. Consumer‑rights groups point out that consent forms were vague, buried in lengthy terms and conditions, and that many users never saw any clear notice about video capture. Moreover, the Indian data‑protection framework, still evolving, does not yet provide a crystal‑clear rulebook for AI training data, leaving a gray area that companies like Pronto seem eager to explore.

Legal experts suggest the episode could become a landmark case for privacy in the age of AI. “If courts interpret ‘implied consent’ loosely, we might see a flood of similar practices across the sector,” says Nisha Patel, a tech‑law attorney based in Mumbai. For now, Pronto has paused the video‑collection effort and says it will redesign its onboarding process to make any data‑gathering explicitly clear.

The incident is a reminder that as artificial intelligence grows more capable, the line between helpful innovation and invasive monitoring can get blurry. Companies will need to balance ambition with transparency, or risk losing the very customers they hope to serve.