Physical AI Is Here. So Are The Data Collection Risks
5/28/202612 min
Physical AI is being seen as the next frontier of artificial intelligence. Not AI that lives on screens. But AI that can navigate and operate in the real world — from humanoid robots and warehouses to factories and homes. But these systems need enormous amounts of real-world human activity data to learn movement and physical tasks. And increasingly, India is emerging as a low-cost training ground for that data collection. In this episode, Host Anirban Chowdhury talks to ET’s Puran Choudhary and Disha Acharya on wearable cameras, AI data pipelines, privacy risks, regulatory gaps and the hidden human layer powering the next AI boom.
Listen in
You can follow Anirban Chowdhury on his social media: X and Linkedin
Check out other interesting episodes like:ET Deep Dive: Swipe Left on Reality,India wants manufacturing at 25% of GDP — will AI in factories help?, Tanay Kothari Wants To Kill The Keyboard,From Doer to Director: The LinkedIn Playbook for the AI Agea, Semaglutide Goes Generic: Big Pharma’s Moat Breaks and much more.
Catch the latest episode of ‘The Morning Brief’ on The Economic Times Online, Spotify, Apple Podcasts, JioSaavn, Amazon Music and Youtube.
See omnystudio.com/listener for privacy information.
Clips
Transcript preview
First 90 secondsAnirban Chaudhuri· Host0:00
[intro music] Physical AI. It's the next frontier of artificial intelligence. [gentle music] Not AI that only writes text or generates images, but AI that can understand and operate in the real world. Think humanoid robots, factory automation, warehouse systems, and even smart home helpers. We're talking real-world tasks, from folding clothes, assembling components, cleaning homes, navigating warehouses, tasks done in autonomous factories and smart environments. But training these systems is much harder than training software AI. Why? Because robots need real-world human activity data to learn movement and coordination. And that is where India is suddenly emerging as this low-cost hub for data collection. In a recent deeply reported ET story, my colleagues Disha Acharya and Puran Chaudhary uncovered how startups, instant house help platforms like SnapIt and Pronto, are collecting physical AI data using wearable cameras and egocentric recording systems. They are recording inside homes, garment factories, and workplaces. So, in this episode, I ask Disha and Puran three key questions.