Investigative Report: Egolab.AI Collecting Indian Factory Workers' Data Without Consent to Train Robots

Summary

An investigative report by Scroll revealed that Egolab.AI, a startup founded in January 2026, has been collecting egocentric video data from Indian factory workers using head-mounted cameras without obtaining proper consent. Workers at Pearl Global Industries' Gurugram facility reported they were asked to wear white head-mounted devices while stitching garments but were not informed of the true purpose of data collection, nor did they provide written or verbal consent. The collected egocentric footage is being aggregated and sold to major AI and robotics companies including Tesla, Boston Dynamics, and Figure AI for training autonomous systems and computer vision models.

Egolab.AI describes itself as "India's largest first-person POV Data Aggregator" and was acquired by Build AI (Build Artificial Intelligence Inc.), a Delaware-registered firm, in March 2026. The parent company now positions itself as "the largest egocentric data collection effort in history." The robotics and AI industry is projected to spend between $1.5 billion and $50 billion over the next 2-3 years gathering 100 million to 1 billion hours of egocentric video data to train robots. Experts warn that much of this data collection is happening without adequate safeguards or transparent consent procedures, revealing a critical vulnerability in the supply chain powering next-generation AI and robotics development.

• The robotics industry expects to invest $1.5-$50 billion on egocentric data collection in 2026-2028, suggesting this practice will scale significantly without regulation
• Data collection infrastructure lacks safeguards and transparent consent procedures, putting vulnerable workers in developing countries at risk as they unknowingly train AI systems

Editorial Opinion

This investigation exposes a critical blind spot in the AI supply chain: the training data for tomorrow's robots is being sourced from vulnerable workers in developing countries with minimal transparency or protections. While demand for egocentric training data is genuine, Egolab.AI's practices show how quickly emerging data markets can circumvent ethical standards when regulatory guardrails are absent. As AI companies race to build advanced robotics and autonomous systems, they bear responsibility for ensuring their data supply chains respect worker consent and rights. This story should catalyze both industry self-regulation and policy intervention to establish baseline standards for labor-sourced AI data collection.