How Did Delhi Police Cameras Identify Criminals Who Were in Jail? Explained ft. Varun | Jist

Our Take

Credit to Jist and Varun for digging into a story that should worry anyone paying attention to automated surveillance, not just folks in Delhi. Police facial recognition tech reportedly flagged people who were already sitting in jail as suspects at a student protest. That's not a minor glitch — it's proof that these systems can manufacture false confidence out of thin air, and that the humans running them often don't question the machine's output before acting on it.

We cover Flock Safety and ALPR networks here in the US, not FRT in India, but the pattern is identical. Automated identification systems get deployed first at protests and politically sensitive gatherings, get sold to the public as neutral tools for catching 'criminals,' and get trusted far more than their error rates justify. When a facial recognition system can't tell someone in a jail cell from someone marching in the street, it's not hard to imagine an ALPR network misreading a plate, flagging the wrong car, or feeding bad data into a police database that follows you around forever.

The lesson isn't 'India has a surveillance problem and we don't.' It's that any government agency running biometric or vehicle surveillance at scale will eventually get it wrong, and the people who get hurt are usually the ones already exercising their rights to speak or organize. Check our camera map to see what's watching your own street, and if you want to push back locally, our take-action page has a place to start.