India's Quiet Entry into the AI Race: How Humans Are Teaching Robots

In a surprising turn of events, India is making a mark in the global AI race, not by developing robots, but by providing the human touch to train them. With its vast workforce and relatively low labor costs, the country has become a hub for companies collecting data to train robots. Meet Tanisha Reddy, a teacher who moonlights as a robot trainer, recording videos of herself performing routine tasks like cooking and cleaning. She's one of the many Indians working behind the scenes to help robots learn. Reddy works for Qanat Consulting Services, a firm that recruits people to record first-person videos, which are then used to train robots. The company's founder, Thaslim Pattan, says they get contracts mostly from companies in the US and China. As the demand for robots grows, so does the need for data collection. Global brokerages like Barclays and Morgan Stanley are bullish on the humanoid robot market, expecting it to reach $200 billion in less than a decade and surpass $5 trillion by 2050. However, with the rising number of competitors, contract prices for data collectors are falling. To stay relevant, Indian companies will need to move up the value chain. Startups like Neocambrian AI and Humyn Labs are already working on creating datasets that can help robots manipulate objects and improve their dexterity. Experts believe that India can become a 'human labor marketplace of the world' and play a significant role in developing operating systems for robots. While the country has a long way to go in manufacturing robots, its strengths in IT and data processing can give it an edge in the AI race. As the market for data collection evolves, India will need to adapt and innovate to stay ahead.