Data annotation jobs grow in smaller cities in South India

In short: More AI companies are relying on workers in smaller cities in South India to label text, images, audio, and video, a key step in training AI systems.

What's going on

Data annotation is the behind-the-scenes work that helps AI learn. It means people label examples, like marking what is in a photo, tagging a sentence’s meaning, or checking a short audio clip. It is like putting name tags on thousands of items so a computer can learn the pattern.

Reporting and job listings point to growing annotation hubs across South India, including places like Salem, Hyderabad, Bengaluru, and Vijayawada. Companies often set up these teams outside the biggest metro areas, where it can be cheaper to hire and train large groups of workers. The work is typically done in an office, broken into small tasks, and sometimes requires specific local languages like Marathi or Telugu.

India already has a large annotation workforce. One report cited around 50,000 freelancers working through online platforms, plus about 20,000 full-time workers at third-party firms that provide annotation services. An industry estimate mentioned in the reporting suggests India could reach as many as 1 million full- and part-time annotation workers by 2030.

What to watch

The big question is what these jobs look like as the industry grows. Some listings mention pay around ₹17,000 per month and on-the-job training, but reports also describe the work as low-paid and insecure, with limited protections. As more everyday products use AI, there may be more demand for this work, and more pressure to improve wages and working conditions.

Source: NYTimes