In a post on X, Khera described the development as “extraordinary and deeply disturbing”, saying it raised questions either about the technology being used by the Delhi Police or the process through which people were identified at the protest site.
“This is extraordinary and deeply disturbing. Delhi Police’s facial-recognition system identified at least 25 people accused of murder, rape and attempted murder as being present at the Jantar Mantar protests - even though they were actually sitting in jail,” Khera wrote.
“That leaves only two possibilities, and both are deeply alarming: 1. The technology is dangerously unreliable. The very system Delhi Police is using to identify and potentially target protesters in Delhi and across the country is so flawed that it is identifying people who were not even at the protest as having been there. OR 2. The technology isn’t the problem. The police is. Criminals and thugs were deliberately brought to the protest site to provoke violence, create disorder and malign the students’ cause,” Khera added.
He further wrote: “Either way, this demands answers. If the technology is this unreliable, why is Delhi Police using it to profile protesters? And if the technology is accurate, how did people who were demonstrably sitting in jail end up being identified as protesters at Jantar Mantar? Delhi Police cannot have it both ways. It owes the country an answer.”
The issue relates to the student protests held at Jantar Mantar between July 20 and 26. Delhi Police had told the Supreme Court that its Facial Recognition System (FRS) identified 2,873 people with alleged criminal antecedents at the protest site.
However, an investigation based on police, prison and court records found that at least 25 people among those flagged in the categories of murder, attempted murder and rape were lodged in Delhi jails when the system identified them at Jantar Mantar.
According to the report, the 25 included 17 people facing murder charges, four facing rape charges, including two POCSO cases, and four accused in attempted murder cases. Three were flagged on July 24, 21 on July 25 and one on July 26.
The findings have renewed questions over the accuracy of facial recognition technology when deployed in large crowds.
Delhi Police had earlier defended the use of FRS at Jantar Mantar, saying it was deployed to identify wanted criminals, absconders and history-sheeters and was not intended to target ordinary protesters. Four FRS units were installed around the protest site and linked to police databases.
In an affidavit submitted to the Supreme Court, police said the system had identified 2,402 people through its biometric database, Crime Kundli, and another 471 through criminal records. It also stated that no action would be taken solely on the basis of an FRS match and that field verification was required before any action.
Khera, however, argued that the latest findings leave police facing difficult questions. If the technology produced false matches, he asked why it was being used to profile protesters. If the matches were accurate, he asked how people recorded as being in jail could have appeared at Jantar Mantar.
The controversy also comes amid a legal debate over police surveillance at protests. A PIL before the Delhi High Court has challenged the use of facial recognition and other surveillance technologies at Jantar Mantar, citing concerns involving privacy, free speech and constitutional safeguards.
About The Author
.jpg)
Comment List