Monitoring Torture: SC’s suo motu action on custodial deaths and CCTV camera non-compliance

On September 4, about a week ago, the Supreme Court registered a suo motu PIL (public interest litigation) regarding the non-working CCTV cameras at police stations. This intervention followed a report by Dainik Bhaskar that 11 custodial deaths had taken place over a period of seven to eight months. The media report also detailed factors like not providing CCTV footage in cases of alleged custodial deaths, citing dysfunctional cameras at police stations in the state. A bench of Justices Vikram Nath and Sandeep Mehta directed that a suo motu PIL be registered, titled ‘Lack of functional CCTVs in police stations. While registering the suo motu PIL, the court pointed out that the report by Hindi daily Dainik Bhaskar said that there were 11 deaths in police custody in the last eight months. The report said that the CCTV footage was not being provided, with the police claiming the cameras were faulty or storage was full, and sometimes citing confidentiality.

The concern of the Court stems from the fact that CCTV surveillance in police stations is obligatory, not discretionary, a requirement repeatedly endorsed not only by the Supreme Court but also by various High Courts. To illustrate, in Akhilesh Pandey v. State of Madhya Pradesh (2018), the Madhya Pradesh High Court instructed all police stations within the State to provide audio-visual CCTV installations. It further warned that:

“In future if it is found that in a Police Station an area was left outside the coverage area of CCTV camera, then such lapse shall be considered as Contempt of Court, and action shall be taken against the Superintendent of Police of the said District and also S.H.O. of concerning Police Station for Contempt of Court.”

At the national level, the Supreme Court (SC) in the case of Paramvir Singh Saini v. Baljit Singh (2020) reiterated the previous directions provided in the cases of Shafhi Mohammad v. State of Himachal Pradesh (2018) and Dilip K. Basu v. State of West Bengal and Ors. In 2020, the SC added more elaborate guidelines to the ones laid down earlier. Whereas Shafhi Mohammad and D.K. Basu laid the groundwork of the videography of crime sites and custodial procedures, Paramvir Singh Saini gave it a wider reach, ordering the installation of CCTV cameras in all police stations and central investigating agencies, with strong oversight committees for adherence.

Key findings in Paramvir Singh v. Baljit Singh, 2020

The Supreme Court in Paramvir Singh Saini v. Baljit Singh (2020) passed a historic order squarely targeting custodial torture and killings by directing the installation and proper functioning of CCTV cameras at all police stations in India. Writing for the Bench, Justice Rohinton F. Nariman laid stress on the fact that this was neither an administrative routine nor a formality but a constitutional requirement based on Article 21 (right to life and dignity) and Article 22 (protection against arbitrary arrest and detention). Like other path breaking judgements of the apex court, this judgement too has been rendered ineffective by lacklustre follow up by state governments and implementation by the police bureaucracy.

The ruling ordered CCTV systems to be installed not just in every police station but also in every agency empowered with arrest and interrogation powers, such as the CBI, NIA, ED, NCB, DRI, and SFIO. This broad range was meant so that each custodial interaction site would be under open scrutiny.

The Court explicitly grounded its directions in fundamental rights jurisprudence by stating:

“Since these directions are in furtherance of the fundamental rights of each citizen of India guaranteed Under Article 21 of the Constitution of India, and since nothing substantial has been done in this regard for a period of over 2½ years since our first Order dated 03.04.2018, the Executive/Administrative/police authorities are to implement this Order both in letter and in spirit as soon as possible.”

In doing so, the Court directed that the installation of CCTV cameras was not an administrative recommendation, but rather a constitutional obligation falling under the ambit of Articles 21 and 22.

The Court devised an oversight framework at the State and District level to aid in the implementation of the installation programme. The Court refined the already existing order of the Central Oversight Body, and added two tiers to the system:

Apart from this, the Station House Officer (SHO) of each police station was personally held responsible for ensuring proper camera functioning. In the event of a breakdown, SHOs were asked to report all custodial activity during that period of inoperability to the DLOC, leaving an audit trail to eliminate misuse.

The Court did more than issue broad guidelines: it specified minimum technical standards. Cameras had to be equipped with night vision and audio-video capabilities, supported by a reliable power source (solar/wind, where applicable) and internet connectivity. The data from the cameras had to be stored for at least one year, with an eventual long-term target of 18 months.

Most importantly, the Court directed for clear signage in local languages at all stations, notifying detainees that they were being monitored under CCTV, that recorded footage would be retained, and that they could approach the NHRC, SHRC, or Human Rights Courts to pursue redressal. Despite such a detailed approach, the impact of this judgement has been stymied by both the executive (state governments under whom the subject ‘law and order’ falls) and the police bureaucracy. The Indian police are notorious for not acceding to any corrective measures to ensure transparency and accountability.

Judicial History: From D.K. Basu to Shafhi Mohammad

The Supreme Court’s directive in Paramvir Singh Saini (2020) did not come out of thin air. It is founded upon decades of jurisprudence that acknowledges custodial environments as a constitutional space within which human rights need to be safeguarded.

D.K. Basu v. State of West Bengal

In Dilip K. Basu v. State of West Bengal & Ors. (2015), the Supreme Court re-examined the question of custodial torture and deaths almost three decades following the initial D.K. Basu guidelines of 1997. In spite of the previous ruling, reports of custodial violence were disturbingly common, and the Court was compelled to tighten measures. The bench of Justices T.S. Thakur and R. Banumathi noted that custodial torture is a “naked violation of human dignity” and constitutes a direct attack on the constitutional right to life under Article 21. The Court underscored that the State cannot claim sovereign immunity in custodial death cases; there is strict liability, and the victim’s families have to be compensated.

The Court gave binding directions for operationalising the protection of human rights. States and Union Territories like Delhi, Arunachal Pradesh, Mizoram, Meghalaya, Tripura, and Nagaland were asked to set up State Human Rights Commissions (SHRCs) in six months’ time, and all vacant positions in SHRCs had to be filled within three months. Sessions Courts were to inform as Human Rights Courts under Section 30 of the Protection of Human Rights Act, 1993. Realising the contribution of technology in accountability, the Court ordered phased installation of CCTV cameras at police stations and demanded prisons to be equipped with cameras in one to two years. It also instructed the non-official guests to be assigned for surprise visits, prosecuting officials responsible for custodial deaths, and at least two women constables being deployed in every police station where women detainees are kept in routine detention.

This judgment was a significant development in India’s custodial jurisprudence, reaffirming the Supreme Court’s watchdog role in ensuring that the basic rights of detainees are safeguarded even in prison. Through the integration of Articles 21 and 22 with institutional mechanisms like SHRCs, Human Rights Courts, and CCTV surveillance, the Court attempted to establish a stronger framework against custodial abuse. This judgment paved the way for subsequent judgments like Paramvir Singh Saini v. Baljit Singh (2020), in which CCTV coverage within police stations and investigation agencies was made compulsory.

 Shafhi Mohammad v. State of Himachal Pradesh

In Shafhi Mohammad v. State of Himachal Pradesh (2018), the Supreme Court considered two important features: the use of videography in criminal investigations and the acceptability of electronic evidence under the Indian Evidence Act. The bench of Justices A.K. Goel and U.U. Lalit placed special emphasis on the fact that videography of crime scenes and recovery operations would significantly enhance transparency, enhance public confidence, and strengthen evidentiary credibility. The Court had looked at international trends, such as body-worn cameras in the US and UK, and noted technology tools were unavoidable for modern criminal justice. To institutionalize this, it directed the Union Government to prepare a roadmap and Standard Operating Procedure (SOP) for videography usage in investigation through consultations with states and expert committees.

The case also explained the law regarding electronic evidence, particularly Section 65B of the Indian Evidence Act. It held that the exclusion of otherwise pertinent and genuine evidence on technical grounds alone would be a denial of justice. Drawing support from the precedents of Ram Singh v. Col. Ram Singh (1985), Tomaso Bruno v. State of U.P. (2015), and Navjot Sandhu (2005), the Court laid emphasis on the fact that electronic evidence could be admitted if authenticity is proved despite the impossibility of adhering to procedural necessities at all times.

The Court designed a systematic five-step programme for the use of videography and CCTV monitoring in criminal investigations. This was to begin with pilot schemes in chosen districts of diverse States and Union Territories for testing infrastructure and an assessment of logistic issues. In the second phase, the accurate Standard Operating Procedures (SOPs) had to be developed which dealt with issues such as camera placement, chain of custody, data storage, and admissibility as evidence under the Indian Evidence Act. The third phase envisioned phased introduction of videography facilities in all the districts in every State to be supported by procurement, allocation of funds, and police training. The fourth stage was aimed at integrating videography into the evidentiary process, particularly ensuring compliance with Section 65B of the Evidence Act, to make electronic records readily admissible in court proceedings. Last but not least, the fifth phase saw the Court instructing the nationwide standardisation and adoption of videography and CCTV practices with oversight arrangements in place at Central, State, and District levels as well as regular auditing and monitoring provisions. Through this measure, the Court aimed at converting technology from a temporary solution into a stable means of preventing custodial abuse and providing a means of improving investigation transparency.

Shafhi Mohammad was therefore a forward-looking development in two senses: it enhanced the role of technology in criminal investigations by mandatory videography, and it relaxed the evidentiary framework by making it possible that technicalities would not stand in the way of true electronic evidence.

Paramvir Singh Saini is the synthesis of these strands: the dignity-and-safeguards paradigm of D.K. Basu joined with the technology-and-evidence model of Shafhi Mohammad. Justice Nariman made this continuity self-explicit, saying that the Saini instructions “follow from previous judgments of this Court, and notably D.K. Basu and Shafhi Mohammad.”

By integrating CCTV surveillance into custodial jurisprudence, the Court aimed to codify constitutional rights into everyday real-time protections, accessible not merely in theory but as an everyday practice at every enforcement office and police station.

Compliance Gaps and the Human Cost of Non-Implementation

Lack of functional CCTV cameras at police stations is not a theoretical bureaucratic failure. It directly affects accountability in custodial fatalities, which continue to be uncomfortably common throughout India. Official statistics provided by the National Human Rights Commission (NHRC) portray a persisting trend: 155 instances of death in police custody were reported in 2021–22 (up to February 28, 2022). Earlier years had the same figures — 100 in 2020–21, 112 in 2019–20, 136 in 2018–19, 146 in 2017–18, and 145 in 2016–17. A total of 669 cases of death in police custody were registered across the country from April 1, 2017, to March 31, 2022. These statistics reveal that even with successive judicial interventions, the magnitude of custodial deaths did not decrease during the last ten years.

The state-level scenario is equally alarming. Rajasthan has alone reported 20 custodial deaths in just two years, based on official revelations. The investigation by Dainik Bhaskar, which led to the Supreme Court’s present suo motu PIL, pointed out that 11 individuals passed away in police custody in Rajasthan for just seven to eight months of 2025 alone, a cohort that graphically captures the cost of defiance of judicial orders. There is no unified national narrative. Some States and UTs (or a few districts thereof) have been credible in their efforts, providing station-wise information, setting up contemporary DVR/NVRs, and forming oversight bodies. A few others failed to submit complete affidavits or filed generic reports without station-level information. The Court docket and subsequent audits yield extremely variable results: whereas a few States have claimed near-universal install rates, independent verification or subsequent press inquiries detect huge numbers of non-recording cameras or insufficient storage. Even the Saini docket itself noted only a handful of States submitting satisfactory action-taken affidavits in the first round.

The causes of the deficit are several and persistent:

1. fiscal inertia — states have not ring-fenced O&M and capital funds even as the Court instructed states to engage finance secretaries at SLOCs;
2. technical impediments — most remote lock-ups do not have power/internet and require custom solutions (solar, local storage);
3. vendor and procurement issues — low quality SLAs and vendors not meeting audio/storage requirements;
4. institutional resistance and low political priority — police leadership at times views CCTV as surveillance of its own ranks and not public accountability; and
5. weak monitoring — SLOCs/DLOCs tend to meet irregularly and do not release independent audits. These reasons account for why a ‘installed/not installed’ binary is deceptive. Where there are missing or low-quality recordings, investigations suffer two simultaneous injuries: victims and families lose an objective contemporaneous account; investigators/courts lose potentially determinative evidence.

This not merely impedes criminal responsibility for custodial brutality but also undermines deterrence: if wrongdoing is unlikely to be recorded on credible footage, the threat imposed by judicial directives is reduced. The Rajasthan cluster illustrates this danger: clusters of fatalities in a short time require taped proof for open-ended investigations; without that, public trust disintegrates.

These figures point to the price of treating Paramvir Singh Saini as a paper directive. Functional CCTV surveillance, with sound, infrared vision, and good data recording, might be able to offer impartial evidence of custodial treatment, deterring torture and providing courts with indispensable evidence. Without that, grieving families are left with little more than bureaucratic assertions of “natural causes” or “suicide,” excuses too frequently unsubstantiated by independent evidence. The continuation of custodial deaths, even in the face of the Court’s painstaking 2020 order, shows that failures in implementation are not simply bureaucratic errors but continuous violations of Article 21’s promise of life and dignity and Article 22’s protection against arbitrary arrest and detention.

Conclusion: Supervisory Jurisdiction and the Path Forward

The order of the Supreme Court to put on record a suo motu PIL regarding custodial deaths and non-functional CCTV cameras is a reaffirmation of its ongoing supervisory jurisdiction over the implementation of fundamental rights. This jurisdiction, though not statutory, is well established in the Court’s constitutional obligation to safeguard life and liberty under Article 21, and the protection of procedures under Article 22. In Paramvir Singh Saini, the Court had already indicated that directives on CCTV emanate out of these constitutional protections. By reopening the case in 2025 through suo motu action, the Court is indicating that compliance is not voluntary — repeated non-implementation is a repeated constitutional wrong.

This form of judicial monitoring is not new. In D.K. Basu, the Court gave out extensive custodial directions and followed through on monitoring compliance for years. In Prakash Singh, it practiced “continuing mandamus” over police reform, calling States back repeatedly to order. The current suo motu action is an extension of this practice that addresses systemic violations of rights as continuous violations in need of judicial oversight, rather than single orders.

However, the repetition of failures also shows the boundaries of judicial instructions with no structural follow-through. For CCTV instructions to become valuable guarantees instead of paper commitments, there must be more robust institutional controls:

1. Independent Technical Audits: Periodic certification of camera operation, storage, and sound-picture quality by accredited independent third parties, with findings publicly disclosed.
2. Station-Wise Public Dashboards: Oversight Committees must release district-wise compliance dashboards — no. of stations, no. of cameras installed, operational status, date of last inspection — to bring in transparency and external accountability. These must be available physically and online.
3. Conditional Funding: Central and State police modernisation funds need to be released only after confirmed compliance, so budgetary inertia cannot be an alibi.
4. Automatic Access Protocols: Victims, families, lawyers, and human-rights groups ought to have rights to obtain and examine footage, within time limits, under a secure chain of custody.
5. Graduated Sanctions: Lack of compliance should invoke departmental penalties for SHOs and SPs, financial responsibility for States, and in cases of wilful disobedience, contempt proceedings against erring officials.

The Supreme Court suo motu PIL thus plays a twofold role: it addresses the immediate crisis of custodial death, and it reaffirms the principle that constitutional rights do not end at the police station door. But whether the long-term viability of this intervention remains dependent on the Court now shifting from generic directions to binding compliance protocols, subordinating technological vigilance into the routine functioning of the police. Will the court, this time, go a step further, in directing structural accountability like monitoring and reporting on the installation of CCTVs for example?

In so doing, the Court can bring Paramvir Singh Saini from a blueprint to a reality, making sure that protection from abuse under surveillance is not an ephemeral judicial experiment but an enduring component of India’s human rights based custodial apparatus.

(The legal research team of CJP consists of lawyers and interns; this legal resource has been worked on by Preksha Bothara)

Related

SC order on CCTVs in Police Stations: Firm stand against Custodial Torture’

Custody, Camaraderie, and Cover-Up: Supreme Court transfers custodial death probe to CBI, slams MP police for “shielding their own”

Custodial Horror in Sivagangai: Ajith Kumar’s death raises chilling parallels with Sathankulam

Death in Chains: Indian Courts award reparation for deaths in custody, deaths rise alarmingly