Digital Tear Gas: Inside the Surveillance State at Jantar Mantar

The Cockroach Janta Party (CJP) student protest erupted in June 2026 at Jantar Mantar, New Delhi driven by widespread outrage over massive paper leaks and structural irregularities in the NEET-UG examination. For around thirty-seven days, student activists maintained a continuous demonstration demanding systemic testing reforms and accountability from top officials. Most significantly, they demanded the resignation of former Union Education Minister, Dharmendra Pradhan. The CJP called for its major “Sansad Chalo” (March to Parliament) on July 9, 2026, scheduling the demonstrations to take place on July 20, 2026.

On the night of 19th July, protestors across India prepared for the worst as Instagram feeds were filled with precautionary measures against tear gas, lathi charge, network jam, and surveillance tech. Many creators shared advisory carousels on Instagram about precautions to take before tomorrow, following which I downloaded Bitchat to my phone to stay connected with my friends in case of an internet jam. Bitchat is a decentralised, peer-to-peer messaging app that works without cellular data, Wi-Fi, or central servers by utilising Bluetooth Low Energy (BLE) mesh networks and the decentralised Nostr protocol.

On the afternoon of 20th July, marked by police brutality, tear gas, lathi charge, stun shocks and even pellet guns — I opened Bitchat to find two of my friends, lost in an endless crowd of almost one lakh people. Anonymous strangers on the global mesh chat looked for their friends and family members, “Bhaiya aap kahan ho?” (Brother, where are you?). Another anonymous message read — “SBI Board ke paas aao” (Come near the SBI Board). Entering the protesting masses meant complete disconnection from the “outside world” —I told my family that I’ll only be able to contact them by the evening and not to worry about me. That noon, reels of people bleeding, police slapping a woman, and tear gas being fired towards unarmed citizens spread across the internet. However, we were disconnected from all this - on that afternoon the only objective was to survive. 

The Internet Freedom Foundation (IFF) in its statement mentioned that no suspension order was made public before the network jam on 20th July in parts of Central Delhi, which deeply affected communications at the ‘Sansad Chalo’ march called by the Cockroach Janta Party (CJP). “Cutting off communications prevents people from reaching out for help and manufacturer panic in a moment which demands calm,” IFF argued the shutdown of network disrupted journalists filing live reports, vendors unable to process UPI payments, gig workers losing a day’s earnings and parents unable to confirm their child’s safety.

According to Times of India, on July 24, 2026, the Indian government directed GitHub to take down Bitchat, citing concerns that it could be used by terrorists, organised crime groups and anti-national elements to evade lawful surveillance. An app used by protesters to track their friends in the endless crowds of Jantar Mantar now poses a “national risk” to the government. 

On the same day, Pragya Singh from The Wire reported how Gujarat police showed up to the house of Shrey Gupta, an Ahmedabad engineer who built a website called The Unbadged to document the 20 July violence. At the police station, Gupta deleted his posts on Instagram, calling for people to submit documented videos of police violence. After seven hours of questioning, paperwork, and documentation of personal details, Gupta was allowed to go. “If something happens to me, I want someone else to carry this forward,” he said to The Wire.

On 22nd July, the Internet Freedom Foundation condemned the restriction and shadow bans of protest-related content on Instagram and demanded transparency from Meta. Protest documentation across Delhi is being marked as “sensitive content,” and in several instances the “see reel” option does not function at all. As I write this piece, the account @free.umarkhalid has also been blocked in India following a government request under Section 79(3)(b) of the Information Technology Act.

On the afternoon of 23rd July, a group of students giving a mathematics exam in class got an urgent notification on their phones, urging them to leave a student-run WhatsApp group created to share protest updates, coordinate together for safety, and for protesters to find people to go with  —the group had been doxxed by an anonymous member on Twitter. “This happened  between two of my quizzes, so I was in a major panic,” recounts Hrishikesh*, one of the members of the group while explaining the doxxing. Hrishikesh* found a Google Drive link on the tweet with screen recordings of the WhatsApp group’s chat and group description - noticing his own name in the video. Within minutes, a WhatsApp group of four-hundred and fifty  people was emptied and abandoned. Students traced the IP address of the doxxer and figured that the anonymous person had used a VPN and a fake Gmail account to share the post. After hours, the students regrouped in another group chat with stricter entry procedures, “we created another group and decided to be a lot more stricter” said Hrishikesh*. Daksh* connected the incident to broader fear of authorities — including possible repercussions from the college administration itself . “I was more hesitant as to what authorities closer to home on campus would do compared to the outside authorities.” Protesters were in fear of their social media groups being tracked, doxxed, shut down and worse, being reported to the authorities. It seemed like the latest example of a pattern playing out across the city — digital protest infrastructure being dismantled from the inside. 

Even after 20th July, the only way of communication between protesters and the outside world was by stepping far from the protest zone or by SMS, which was a choppy method — however it was the only respite for many on days like the 20th.

However, it is worth noting that the watchful eye of Delhi police doesn’t stop at digital surveillance on social media and network jamming to block communication within Jantar Mantar. Brut. reported that a petition filed by Jawaharlal Nehru University Students' Union (JNUSU) former president Aishe Ghosh alleged that since 20 June, Delhi police have maintained constant surveillance at Jantar Mantar through a permanent observation tower, photographers, videographers and AI-enabled monitoring equipment. The Center responded by defending the videography of protesters and contended that “every protest is videographed” and that a “claim of privacy in a public place is ironical,” reports Thapliyal from LiveLaw. According to a report by NDTV, on 28 July, the CPI(M) alleged that the Delhi police attempted to arrest Aishe Ghosh at the party office.

A post by @watchthestate on Instagram reported that police personnel used AI-enabled control rooms to continuously identify demonstrators — many reels by those on the ground also showed police using Meta glasses to record the crowd. Watchthestate also highlighted that “predictive policing” can reinforce discrimination by directing more arrests towards marginalized communities.

The patterns of surveillance in the Jantar Mantar protests reflect how repression has also moved from analogue to the digital —placing marginalised communities at the center of systemic targeting. On the night of 22nd July, a drone flew above us in the sky between the protesters and the police at the Jantar Mantar intersection. The Indian Express reported how the Ikshana vehicle used face detection and recognition, number-plate recognition, traffic management, crowd monitoring, and on-the-spot video analytics. The system detected thousands of faces and live-updated them to police databases. 

David Garland’s The Culture of Control (2001) describes how protests and “deviance” are framed narrowly as management problems rather than a democratic right to dissent. Garland argues that two criminologies act simultaneously when it comes to demonstration surveillance; the police a) singles out and targets presumed “bad” protesters/groups for special surveillance, and b) applies blanket preventive filming to everyone at a demonstration, "bad" or not, just because being in that situation makes surveillance economically rational from a control standpoint.

Protesters ask difficult questions rather than settling for the answer that surveillance is used to “maintain law and order” — who is the surveillance serving? Who owns and accesses the data? Will the Delhi Police use these records to identify the police officers who unleashed violence on the protesters? Will CCTV recordings help provide justice to the women who got sexually harassed at the protest? 

The patterns of surveillance have not been specific to the Jantar Mantar protests of July 2026. According to Trivedi’s report for The Hindu, Delhi Police used Automated Facial Recognition System (AFRS), originally acquired in 2018 to trace missing children — to screen protest crowds and identify "habitual protesters" and "rowdy elements” at the Anti-CAA/NRC Protests in 2019-2020. At the Farmers’ Protest (2020–21 and 2025), the Centre ordered suspension of internet services at Delhi's border points. Human Rights Watch documented Indian authorities using threats, excessive force, and internet shutdowns to stop farmers from holding peaceful protests. Farmers dealt with drones firing tear gas shells, facial recognition systems, and having their phones tracked, in addition to internet shutdowns and suspension of their accounts on X, as suggested through comprehensive reports by Amnesty International and Human Rights Watch.

At the same time, activists have responded through legal challenges and through “counter-surveillance” – filming the police themselves. Instagram feeds over the past few days have been flooded with edits and first-person videos of police violence. Brut. also published video footage documenting an officer sexually assaulting a woman protester. As the graffitied walls of Jantar Mantar were painted over by the New Delhi Municipal Council (NDMC) on the night of Dharmendra Pradhan’s resignation, the first person documentations of the protest and police brutality on social media platforms like Instagram will go on to be a part of the public memory. 

At the same time, many creator accounts and posts on Instagram have also been blocked, restricted and ‘shadowbanned’. This has been happening amid a Delhi Police operation in which social media platforms are being told to remove posts carrying purportedly objectionable content about the Prime Minister, reported Rout from Newslaundry. Journalist Suruchi Kumar from The Hindu interviewed several women facing doxxing and calls from strangers to “teach them a lesson” for posting about the protest — several posts were reshared by meme pages and social media accounts calling for their arrests. Such an environment continues to make it seem difficult and dangerous to even talk about the protest on social media, especially as a woman.

Amid surveillance and fear of detainment activists, volunteers and protesters continued to mobilise thousands of people to join the protests by sharing live updates through photos and WhatsApp forwards (wherever internet was available). Many moved away from Meta apps to more decentralised software to communicate and mobilise resources. People across the nation extended their solidarity on the night of 20 July by sending free food through BlinkIt, Swiggy and Zomato — so much that CJP had to urge supporters to not send food any more. Quick commerce apps bridged the gap between digital solidarity and in-person support. Thousands who could only watch the helpless visuals of students getting beaten up found another way to act — a small order made sure a protester did not sleep hungry or injured. In a few instances, delivery personnel were stopped by police officers near protest zones with food being confiscated or riders being temporarily detained and questioned —an interview of a food delivery partner uploaded by @thecaravanmagazine on Instagram suggests this.

At the same time, many volunteers who were donating food through crowdfunding hid their identity to avoid risks of detainment and being incriminated. I spoke to Shourya*, who along with his friends, raised money through social media to directly provide food and healthcare supplies to Jantar Mantar. Even while providing essential goods for the protest, Shourya* and his friends took exceptional measures to avoid surveillance. “We did not put QR codes on our stories,” he recounted, “we had crafted a message which minimised the banking details we gave to the person, but also made sure there’s transparency on who is donating and who is receiving it at the end of the day” since payment interfaces like UPI are easily trackable. Alongside that, a lot of the purchases made by the volunteer group were made offline and from far away hospitals, Blinkit delivered resources far away from the site, “even for cabs, we would have meetup points away from the protest site,” Shourya* said. Attempting to counter surveillance at multiple levels, Shourya* and his friends made sure to even be cognizant of their outfit, “every person who was donating, including myself —we wore a mask at all times and monochrome clothes”.

The past week at Jantar Mantar has looked like surveillance as an entire infrastructure: network jams, AI-enabled control rooms, facial recognition vans, drones, doxxing, and platform takedowns, working together the way they have before, in the Anti-CAA-NRC protests and the Farmers' movement. But the same week also had Bitchat helping friends find each other in a crowd of a lakh people, even as the government moved to have it taken down. A student built a website to document police violence. Reels and edits of police brutality kept circulating faster than platforms could label them "sensitive." Delivery riders kept protesters fed even when it meant getting stopped and questioned themselves. A group of volunteers continued assisting protesters through crowdfunded healthcare and food items. None of this cancels out the fear or the injuries or the accounts that were quietly erased. The questions raised here, about who owns this data, who it protects, and who it's used against, aren't going away once the tear gas clears. If anything, Jantar Mantar has only made them harder to ignore.

*Pseudonyms have been used to protect the identity of the individuals.

[Edited by Maya Ribeiro and Avika Mantri]