Delhi Police Defends Facial Recognition Use at Protests
Delhi Police told the Supreme Court it used facial recognition technology only to identify individuals with criminal records, not to surveil the entire crowd. Petitioners disputed the claim.

Delhi Police informed the Supreme Court on August 18 that its use of facial recognition technology (FRT) at recent protests was limited to identifying individuals with existing criminal records, not for general crowd surveillance. Lawyers for the petitioners contested this assertion, arguing that the technology inherently captures all faces before any matching can occur.
The court heard arguments regarding surveillance during protests at Jantar Mantar. Experts explain that facial recognition systems first detect a face, extract its unique features into a template, and then compare this template against a database. This process, often described as "one-to-many" identification, requires the detection and processing of all visible faces before a potential match can be made against a criminal database.
In its affidavit, Delhi Police stated that the FRT system only captured faces of individuals previously booked for offenses and did not automatically profile everyone present. The system reportedly contained data only on those accused of serious offenses, excluding minor infractions. The police reported identifying 2,873 individuals with criminal records during the protest period, with 92 linked to multiple cases. Solicitor General Tushar Mehta argued that the system does not indiscriminately identify everyone but checks detected faces against the National Crime Records Bureau's database of individuals with serious criminal antecedents.
However, petitioners' counsel countered that the technology inherently collects all images and that the mere collection and processing of such data raise constitutional concerns. Questions were also raised about the source of the underlying database and the fate of faceprints collected from individuals not on criminal records. India currently lacks specific legislation governing biometric or facial recognition surveillance, and past investigations have revealed deficiencies in oversight and privacy assessments related to the technology's deployment.