Digital Policing in Pakistan: How Safe City Technologies Expand Surveillance Without Safeguards

Author: Sheikh Sibghat Ullah

March 3, 2026

[Image Source: Pakistan Today, “Digital Vigilance: ICT-15 in Action” (April 12, 2024).]

Framing the Problem: From Crime Control to Permanent Surveillance

In Pakistan, the Safe City projects have, in the last decade, not only ceased to be experimental security projects but have become an infrastructural backbone to urban policing in Lahore, Islamabad, and, more recently, Karachi. What was initially an anti-terrorism measure against violence has quietly become an architecture of daily surveillance, with thousands of networked surveillance cameras, biometric interfaces and central command and control rooms becoming a standard part of the administration of the public sphere. In this case, the subject of surveillance is no longer a suspect but the population itself. Facial recognition links national ID databases with real-time tracking, which makes it possible to identify a person and map their movements at any moment in time, thereby obliterating the difference between a targeted investigation and general surveillance of the population.

These systems are still mainly discussed in the context of urban safety and crime prevention. Concerns about security have, to a great extent, sidelined the general discourse on privacy, dignity, and due process, a shift that is subtle, but which has far-reaching implications. Technological capacity has already surpassed legal restraint, turning surveillance into a normal form of state power. As a result, the system largely functions without any actual legal or constitutional challenge.

Applicable Legal Framework: What the Law Requires

On the constitutional level, the growth of digital monitoring has a direct interaction with three guarantees of the Constitution of Pakistan that are interconnected to one another. Article 14 of the Constitution  guarantees personal dignity and privacy as a fundamental right of each citizen. As per the Article, the inviolability of dignity is not limited to the domestic sphere but also where a person reasonably expects privacy. The principle of personal liberty enshrined in Article 9 goes further to restrict the authority of the state to subject individuals to perpetual surveillance unless in accordance with the law. Article 10A ensures a fair trial and due process. This establishes strict regulations concerning the collection, storage and use of evidence. These considerations become particularly significant when considering the application of computer algorithms and biometric data to help advance police investigations.

These three domestic guarantees cohere with the International Covenant on Civil and Political Rights (ICCPR) that binds Pakistan. Article 17 prohibits any illegal interference with privacy, whereas Articles 19 and 21 guarantee the right to speak freely and peacefully assemble, which is easily undermined due to the presence of constant surveillance. International jurisprudence has always demanded that any encroachment into these domains would have to meet the cumulative requirements of legality, necessity, proportionality and pursuit of a legitimate goal. Legally, surveillance can only be legitimate under three conditions: the law permitting it must be clear, it must be pre-authorised by a judge or other independent body, and there must be a mechanism to assist any person subjected to illegal surveillance.

The Regulatory Gap: How Safe City Operates Outside These Safeguards

Although surveillance is regulated by the constitutional and international standards, Pakistan lacks a unified law on data protection or an independent body to oversee the process of data collection and utilisation. Proposed data protection bills have been in circulation over the years, but practically, there is no legal framework in place on how biometric data, facial images, or location history collected under the Safe City systems could be stored, shared, or challenged. The Safe City Authorities themselves act primarily according to administrative laws and executive orders. Although these structures provide the frameworks under which monitoring is carried out, they do not contain the legal regulations that define when and how intrusive surveillance may lawfully be carried out.

This gap in the law is apparent, as there is neither a system of warrants nor a clear set of regulations that permit the use of advanced surveillance tools. No publicly available guidelines exist on when facial recognition is allowed, the extent to which one can retrieve the past movements of an individual, or when protests and political gatherings are legally allowed to be tracked. These choices are left to police discretion, are not pre-authorised by a court, and do not receive any ex-post review.

Simultaneously, the core technologies are provided and frequently operated by commercial vendors. Their algorithms are not transparent and cannot be subject to legal examination, despite undertaking activities that directly influence arrest, investigation, and identification. Civilians have no way to trace their own data; they have no way to know how long their data is stored, who it is shared with or how it will be used against them in future due to the lack of effective parliamentary and judicial checks and balances.

Human Rights Impacts: Who Is Affected and How

Basic civil and political rights are directly affected by the development of digital policing. The ubiquitous presence of cameras, facial recognition, and tracking of the location contributes to an environment where it is no longer anonymous to attend protests and political events. This has a chilling effect on freedom of assembly since the people know that they can be tracked with the cameras, identified, and further scrutinised later. To people such as journalists and lawyers, constant surveillance kills freedom of speech. It develops a culture of self-censorship where individuals prefer to remain silent rather than voice an opinion. 

Function creep aggravates these risks. Technologies presented under the pretext of combating terrorism and serious crimes are being employed more often in the process of managing public order and social policing.Once developed to trace dangerous suspects, it is now being directed against activists and political networks. It is now used to track the routine movements of ordinary people in their respective communities. This movement blurs the distinction between security policing and political surveillance.

The effects of this tracking are unequally distributed. During demonstrations, political activists and opposition movements are more likely to be identified and tracked. Biometric and predictive systems might be harmful to religious and ethnic minorities who are already overpoliced, as they tend to reflect the biases of their creators, and thus, undeserved targeting becomes even more probable. The urban poor, living in densely monitored areas, are also disproportionately visible to the surveillance apparatus.

Finally, individuals lack effective remedies. There is no practical way to know whether one has been subjected to facial recognition or movement tracking, to challenge the legality of such monitoring, or to seek correction when surveillance-based decisions are wrong or discriminatory.

Case Illustration: Safe City Authorities in Practice

In Pakistan, the idea of Safe City systems is not merely an abstract concept, but an active system of surveillance infrastructure, which determines routine policing as well as event-specific monitoring. The Safe City Authority in Lahore has started the modernisation of approximately 8,000 cameras with facial recognition and automatic number plate recognition (ANPR) technology that will not only allow law enforcement agencies to reveal traffic offenders but also monitor people and vehicles in real time. On the same note, the Sindh Safe Cities Authority has also implemented facial recognition that is connected to criminal databases, whereby any match will result in alerts sent to the police in a few seconds.

These systems are connected to central command and control centres, where the video feeds, vehicle data, and biometric alerts are being monitored constantly. The Sindh Safe Cities Authority is required to establish an integrated platform for such real-time coordination. In Islamabad and other cities, tens of thousands of surveillance activities have been managed by the command rooms, including monitoring during protests, religious processions, and major public events. When attached to the national identity system, surveillance of the population increases even further since face and vehicle surveillance can be cross-referenced with central government data in real-time. This wide-scale connection brings about the possibility of mass identification with no distinct legal boundaries.

Reform Agenda: Principles for Rights-Respecting Digital Policing

The primary reform agenda in Pakistan should be a strong law on data protection and surveillance that spells out permissible use of personal data, establishes restrictions on what can be collected and stored, and establishes enforceable rights of citizens. The Ministry of Information Technology and Telecommunication has published the Personal Data Protection Bill 2023, which suggests the regulation of the processing, sharing, and transfer of personal data and registration and breach notification requirements. However, according to legal commentators, this draft requires more robust human rights protection and wider consultation.

Even with the existence of a data protection law, intrusive technologies, including facial recognition and real-time tracking, must not be enabled without the approval of the courts. The international standards on surveillance underline that the violation of the right to privacy must be laid down by the law; surveillance must have a legitimate purpose and must be necessary and proportionate to that purpose.

Compliance must be monitored by an independent oversight body, such as a national data protection authority that has audit and enforcement powers to investigate and impose penalties in case of violation. The draft bill envisions the establishment of a National Commission for Personal Data Protection, which needs enhanced independence and enforcement power.

Transparency requirements must also be legislated. This set of standards would encompass all aspects of data storage to the making of decisions by algorithms, and most importantly, it would compel authorities to consider the implications of surveillance tools before their implementation in the community. The principles of the Freedom Online Coalition on surveillance governance delineate the way the surveillance policy should be framed using legality, necessity, and transparency.

Lastly, protections should be implemented for political, journalistic, and vulnerable groups. Surveillance devices must never be employed to silence political dissent, and they must also not be employed to attack particular groups or to establish old biases as a permanent component of the system. These values are compatible with more general international human rights principles, which caution that unchecked surveillance may compromise basic liberties.

Conclusion

The high-paced development of Safe City technologies in Pakistan is an indicator of a structural transformation of relation between the state and citizens. Surveillance is not a special reaction to certain threats anymore, but a regular aspect of city management. However, this technological change has not been accompanied by strong legal architecture. The constitutional rights to dignity, privacy, and due process, along with the Pakistani obligation under the ICCPR, demand that any intervention in individual liberty must be based on legislation, necessity, and subject to independent examination. In the absence of definite statutory boundaries, judicial discretion, and effective sanctions, digital policing is likely to become the normalisation of a kind of power that is efficient but unaccountable. The central challenge, therefore, is not whether technology can enhance security, but whether the rule of law can still discipline the ways in which security is pursued.

---

Glossary

• Algorithmic Opacity: The lack of transparency in how automated systems make decisions or identify individuals.

• Biometric Data: Personal data based on physical characteristics, such as facial images, used for identification.

• Chilling Effect: The deterrent impact of surveillance on lawful activities like protest, journalism, and free expression.

• Digital Policing: Use of digital technologies such as CCTV, facial recognition, and databases by law enforcement for monitoring, investigation, and control.

• Due Process: The constitutional requirement that state action, including evidence collection and surveillance, follow fair and lawful procedures.

• Facial Recognition: Biometric technology that identifies individuals by matching their facial features with images stored in state databases.

• Function Creep: The gradual expansion of surveillance technologies from counterterrorism to routine political and social monitoring.

• Independent Oversight: Supervision by an autonomous body to ensure surveillance powers are used lawfully and can be challenged when abused.

• Judicial Authorisation: Prior approval by a court before intrusive surveillance measures are carried out.

• Proportionality: The principle that surveillance must not be more intrusive than necessary to achieve a legitimate security purpose.