Delhi Police's Most-Filmed Officer This Week Was Also Filming Everyone Else
ADCP Sandeep Lamba was removed from CJP protest duty this week after a video showed him slapping a woman near Parliament. A separate video shows him at the same protest site wearing what appear to be smart glasses.
In a video circulating on Instagram and Reddit, a police officer stationed at Jantar Mantar for the Cockroach Janta Party (CJP) protest is seen wearing what appear to be smart glasses — identifiable by a thicker frame and a small embedded lens near the hinge. The officer has been identified as Sandeep Lamba, the Additional Deputy Commissioner of Police (North East District) and the video was originally uploaded on July 22.
Unlike a phone held up to record or a CCTV camera bolted to a pole, the glasses sit on the face and are easy to miss. Smart glasses can capture video, stream footage, and run facial recognition in real time, often with no visible signal that recording is underway.
A clarification worth making up front: some posts and publications have described Lamba's glasses as Meta glasses. That isn’t confirmed. They could just as likely be from AjnaLens, the Indian firm that has previously supplied smart glasses to Delhi Police.
The distinction isn’t just about getting the brand right. Meta’s glasses are a consumer device—something anyone can buy and use. But if these are police-issued, they are part of a law enforcement system, potentially linked to facial recognition and deployed on duty at a protest site.
The privacy risks may overlap, but the source—and the stakes—don’t. This isn’t a bystander’s gadget; it’s state surveillance infrastructure.
Lamba is already the subject of a separate controversy.
On July 20, during the CJP-led march toward Parliament, a 36-second video showed him slapping a woman protester near the Parliament complex. The footage went viral and was raised in the Delhi High Court by senior advocate Gopal Sankaranarayanan, who identified Lamba by name and described the assault as "completely unprovoked".
Lamba has since been removed from protest duty and directed to report back to his district posting. Delhi Police said he "lost cool" in the "spur of the moment".
Surveillance, now on the move
The CJP sit-in at Jantar Mantar — Delhi's designated protest site — has been running since June 20. The movement, founded by political communications strategist Abhijeet Dipke, grew out of Chief Justice Surya Kant's May 2026 remark comparing unemployed youth to "cockroaches."
It has since coalesced around demands for Education Minister Dharmendra Pradhan's resignation over the NEET paper leak scandal, which led to the cancellation of the exam and, in the days that followed, at least a dozen student suicides. Activist Sonam Wangchuk joined, with a three-week hunger strike in late June.
The protest site has been tightly controlled from the start: heavy barricading, a large and continuous police presence, and persistent videography of demonstrators.
On Monday, when protesters attempted to march toward Parliament, police responded with lathi-charge and tear gas. Nearly 180 people were injured, according to Delhi Police — 118 security personnel and 60 protesters.
The smart glasses add a layer to what is already an intensely surveilled space. The containment at Jantar Mantar has been physical — barricades, cordons, force. The glasses suggest it is also becoming something harder to see and harder to challenge.
Not The First Time
Delhi Police tested wearable surveillance earlier this year. In January, ahead of the Republic Day parade, the force announced it would deploy AI-enabled smart glasses developed by AjnaLens, a Mumbai-based startup, along Kartavya Path. Additional Commissioner Devesh Kumar Mahla confirmed that sub-inspectors would wear the devices at entry gates, junctions, and along the parade route.
The glasses were designed as mobile extensions of the police's facial recognition system.
As officers move through a crowd, the device scans faces in real time and matches them against a criminal database — with the number of records in that database reported variously as 10,000 to 65,000, depending on the source. Reportedly, the smart glasses flag individuals with colour-coded markers during scans—green boxes indicate no criminal record, while red boxes signal a match in criminal databases.
Matches above a confidence threshold trigger alerts on the officer's phone. The glasses are also equipped with thermal imaging to detect concealed weapons or metallic objects.
One detail from the Republic Day briefing stands out: police officials said the system could attempt to "update" old images, using AI to generate current likenesses from photographs that are decades old.
How the glasses actually performed on Republic Day is unknown. When MediaNama filed a Right to Information request seeking records of the deployment — approvals, communications from the Home Ministry, tender documents, procurement details — Delhi Police replied that it had no records, no information, and no tender copies related to the glasses.
Apart from the January 22 press conference, no public evidence has confirmed the deployment took place or that it caught any suspects.
What has already gone wrong
What happens when wearable cameras meet protest spaces is not a hypothetical. It has already played out — at the same protest site.
In March 2026, at a protest against the Transgender (Protection of Rights) Amendment Bill at Jantar Mantar, a transfeminine graphic designer named Shubnam was filmed without their knowledge using Meta AI Ray-Ban smart glasses.
The person recording was a content creator who goes by @parley.96 on Instagram, who specialises in ragebait pranks. The resulting video was edited, uploaded, and propelled by Instagram's algorithm to over 74 million views across multiple uploads before the platform took it down on June 18 — more than a month after it was first posted.
Decode broke this story, speaking to Shubnam directly about the fallout. What followed was a wave of transphobic harassment: memes, AI-generated images, coordinated pile-ons from right-wing accounts on X.
Shubnam, who had spent years carefully navigating their visibility, told Decode: "All it took was one unconsented video to rewind two decades of progress that I made all on my own."
That incident and this one both happened at Jantar Mantar, and both involve smart glasses recording protesters without consent. But they sit on opposite ends of the same problem.
If Lamba, however, is wearing police-issued glasses built for institutional facial recognition, that marks a different order of surveillance. One points to a company failing to build in meaningful consent. The other is about the state’s own surveillance capacity showing up, uninvited, at a protest it is also policing.
The protest had posted volunteers to discourage filming, precisely because many people present were not publicly out. The glasses bypassed that entirely.
When police hold the camera
When the state is the one wearing the glasses, the problem scales.
During the 2020 Delhi riots, police used facial recognition to identify suspects from crowd footage and match them against government databases. Home Minister Amit Shah told the Rajya Sabha that 1,922 perpetrators had been identified through facial recognition software — roughly 73 percent of the 2,619 people arrested in connection with the riots.
But a 2025 investigation by The Wire and the Pulitzer Center found that some of those individuals were arrested solely on the basis of facial recognition matches, with no corroborating evidence or credible witness testimony.
One man, identified as "Ali" in the investigation, was arrested from the alleys of Chand Bagh and spent more than four and a half years in pre-trial incarceration. Media reports indicate that more than 80 percent of the riot cases heard so far have resulted in acquittals or discharges.
Delhi Police's own record on transparency around these systems is thin. When asked through RTI about whether facial recognition was used in investigating the riot cases, the police initially denied the information on intellectual property grounds. The Central Information Commission called their replies "patently incorrect" and noted they showed "no application of mind".
Smart glasses take the same capability — scanning faces, matching them against databases, flagging suspects — and make it mobile. What was once a camera mounted on a pole, or a van stationed at a protest perimeter, can now walk through the crowd.
The legal gap
India has no law that specifically governs the use of facial recognition technology by police, let alone wearable surveillance devices in public spaces.
The Digital Personal Data Protection Act of 2023 sets rules for how personal data, including biometric information, should be processed — but it was not designed to regulate real-time facial recognition, and it grants broad exemptions to the state for "public interest" and "security" purposes.
A private member's bill — the Facial Recognition Technology (Regulation of Police Powers) Bill, 2023 — has been pending in the Rajya Sabha since it was introduced, proposing magistrate-level authorisation for police use of FRT. Private member's bills rarely pass in India.
While India’s policing frameworks already incorporate facial recognition in investigations, with the 2024 Bureau of Police Research and Development compendium directing officers to use it to identify suspects in riot cases where images are available, there is no publicly available SOP governing wearable surveillance like smart glasses.
The questions pile up without a framework to answer them. Who has access to the footage captured by the glasses? How long is it stored? Is there a review mechanism? Can individuals challenge a match or request deletion? When the glasses are worn at a protest — a constitutionally protected activity — what rules govern their use?
Right now, none that are publicly known.
At Jantar Mantar, the CJP protesters are still there. But the terms of being present have changed. You are no longer just being watched. You may be being scanned, matched, and archived — and you will likely never know when it started.