Tata Elxsi's New AI Watches Factory Cameras So Workers Stay Safe

Most factories, warehouses, and chemical plants have hundreds of security cameras. But here's the awkward truth: almost nobody is actually watching them. They record footage, and safety teams only review it after an accident has already happened. One control-room operator cannot realistically keep an eye on hundreds of live feeds at once, so spotting an unsafe situation can take 15 to 45 minutes — long enough for someone to get hurt.

Tata Elxsi, an Indian engineering and design firm, built a fix called IRIS. It's a system that uses computer vision — basically, AI that understands what's happening in a picture or video — to watch those camera feeds automatically, all day and all night. When it sees something risky, like a worker in a hazardous zone or a vehicle in the wrong place, it sends an alert within seconds. It doesn't replace safety programs or human safety officers. It gives them a second pair of eyes that never blinks.

A clever part is where the work happens. Instead of streaming every second of video to the cloud, which would be expensive and slow, IRIS does the first round of analysis right there at the factory on a small computer. It only sends the relevant frames and details to Amazon's cloud service. That keeps costs down, protects privacy a bit, and makes the whole thing faster. Tata Elxsi even chose a data center in Mumbai so information stays in India, which matters for local regulations.

The practical payoff is scale. Whether a company runs five cameras or five thousand, it can add more without hiring more people to stare at screens. The same approach could eventually show up in offices, hospitals, construction sites, or anywhere cameras already exist but nobody has time to watch them.