The Number Behind the SC's FIR Order: 2,738 and the Test It's Forcing the Court to Build

The Real Work in This Order Isn't "FIRs Can Be Withdrawn"

• It's Building a Test to Sort 2,738 Faces From a Crowd Article 19(1)(b) protects the right to assemble peaceably and without arms, subject to reasonable restrictions under Article 19(3) for public order. That's the easy part of this case and it isn't really what the Supreme Court spent Monday on. The harder question, forced on the Court by the Solicitor General's own disclosure, is this: once facial recognition technology tells the state that 2,738 people with records of murder, rape or child abuse were present at a protest site alongside thousands of genuine student demonstrators, what test do you use to separate the two - and does that test survive contact with Article 21's privacy protections? The Court's answer, so far, is a classification: "criminal antecedents," the phrase that had earlier been used to deny interim protection to some protesters, now means only "grave and heinous offences." That's not a small clarification - it's the Court quietly building a workable filter. A minor prior offence or no record at all, puts you in the protected category, eligible for FIR withdrawal under Section 360 of the Bharatiya Nagarik Suraksha Sanhita. A record involving murder or a POCSO case keeps you outside that protection entirely. The test is doing real classificatory work precisely because the alternative - either shielding everyone at the protest site or coercively pursuing everyone - was never workable once facial recognition had already sorted the crowd for the government. What makes this genuinely unsettled, not just procedurally tidy, is what N. Hariharan raised for the petitioners: the only way this sorting was possible at all was mass biometric scanning of a protest crowd without consent, an issue the Puttaswamy judgment's privacy protections speak to directly but haven't yet been tested against. The Court has not ruled on that question - it has deferred it, choosing instead to fix the downstream classification problem (who gets protection) while leaving the upstream method (how the state identified who's who) for August 18. That sequencing is the real precedent to watch. If facial recognition-derived classification becomes an accepted basis for who does and doesn't get FIR withdrawal in future mass protests, the Court will have effectively legitimised the tool by using its output - even while the tool's own constitutionality remains formally unresolved. For any GS2 answer on this, the exam-relevant insight isn't "the right to protest was upheld." It's that BNSS-era criminal procedure is now being tested against a policing method the Constitution has never directly addressed and the classification the Court builds here - grave and heinous versus everything else - may end up mattering far beyond this one protest.