A Protest That Ended, A Dataset That Didn't
Even now, in a week where the actual grievance has already been resolved, dear reader, the surveillance apparatus built to manage it is still running. The protest was about a leaked exam. Students organizing under the banner of the online collective Cockroach Janta Party gathered at Jantar Mantar demanding the resignation of the minister overseeing the NEET-UG 2026 paper leak. He resigned on July 25th. The protest ended. What did not end, according to a petition now sitting before the Supreme Court, was the biometric record Delhi Police built of everyone who showed up to watch.
The Toolkit
In a development that will surprise no one who has been paying attention, the response to a few thousand students standing in a public square involved more surveillance infrastructure than most cities deploy for actual manhunts. Two Ikshana mobile units, eight-camera vehicles originally rolled out for the 2023 G20 Summit, parked near the protest site running real-time facial recognition against police criminal databases. Officers on the ground wore AI-enabled smart glasses. And, per the petition, biometric data was pulled through the National Crime Records Bureau's Abhigyan app, a fingerprint tool designed to identify criminal suspects in the field, matched live against the National Automated Fingerprint Identification System.
That system exists, on paper, to identify people already suspected of crimes. Sources within the surveillance-technology community note that pointing it at a permitted protest is less "policing" than "cataloguing," and the distinction matters quite a bit if you're a journalist standing at the edge of a crowd with a notebook and no criminal record to speak of.
The Machine does not know the difference between a suspect and a bystander with press credentials. It was never asked to.
Two Thousand Five Hundred Reasons, None of Them Explained
Police have confirmed the system flagged more than 2,500 people at and around the protest site as having suspected criminal backgrounds, a number officials have offered as proof the deployment worked. What officials have not offered is a breakdown of how many of those 2,500 were actually protesters exercising a constitutional right to assemble, how many were journalists doing their jobs, and how many were just people who happened to be standing near Kerala House on the wrong afternoon. The petition filed under Article 32 by Rajya Sabha MP AA Rahim argues that photographs, fingerprints, and facial scans were collected indiscriminately, from protesters and bystanders alike, with no statutory framework governing any of it. A parallel petition from former student union president Aishe Ghosh is making the same argument to the Delhi High Court.
The government's defense, offered through official channels, is that this was routine — the same class of measure used for Republic Day, Independence Day, and festival crowds, deployed here because of unspecified "intelligence inputs" about cross-border interest in the student movement. One presumes a leaked chemistry exam does not typically draw the attention of foreign intelligence services, but sources confirm the claim was made anyway.
Where This Leaves Us
Nobody disputes that the protest happened, that the minister resigned, or that the cameras were real. What's contested is the part that happens after the crowd goes home: a biometric file, built without a law authorizing it, of everyone who exercised the right to show up and object. And yet — as if this were not enough — the technology used to build that file was justified, in the moment, by the same all-purpose national-security logic that justifies almost anything, applied here to a dispute over standardized test security.
The exam scandal got a resignation. The surveillance scandal, so far, has gotten a court date.
Sources: Bar and Bench — CPI(M) MP moves Supreme Court against facial recognition tech at Jantar Mantar protest · The Tribune — Delhi Police deploy facial recognition system, AI surveillance vehicles at Jantar Mantar protest site
