It’s a striking sight: a large hall with multiple pods, each with four workstations equipped with multiple monitors and other technological equipment. The walls are covered with TV screens, most of which are showing a live feed from different parts of the city. Some of the screens show traffic hotspots, others show faces being automatically detected and matched against a database, while a final set of screens acts as a dashboard of key indicators.
This is the Punjab Police Integrated Command, Control and Communication System, or PPIC3, as police officers refer to it. It’s a sight that greets hundreds of visitors every year: students, politicians, musicians, athletes, and ordinary citizens from all walks of life regularly visit the Punjab Safe Cities Authority (PSCA) headquarters at Qurban Lines on Mall Road, Lahore. They take photos outside before a guided tour of the facility, surrounding a police officer who explains how the Punjab Police are using the network of over 8,000 cameras and other sensors installed in over 1,500 locations across Lahore to fight crime, manage traffic, and deploy emergency services.
These tours, and indeed the Safe City initiative itself, are part of a broader transformation undergoing policing in Pakistan, and represent a substantial investment in rebuilding the trust of citizens in the state and its law enforcement agencies. For a country ravaged by decades of urban violence and terrorist attacks, the introduction of a new security apparatus is meant to be reassuring, representing a significant investment in government capacity. And it’s not just about doing a better job; it’s about changing the way in which government functions as well; in fact, the PSCA website itself proclaims “BEGINNING OF A NEW POLICE CULTURE”. The promise is that citizens will be freed of the bureaucracy, ineffectiveness and corruption of police stations, and can instead build a different kind of relationship with law enforcement authorities. The extensive public relations campaign, which extends beyond tours to enthusiastic posts on social media, as well as radio and TV proclamations of success on a nearly daily basis, has led to support and generous budget allocations from multiple provincial governments to scale the Safe City project from Islamabad to Lahore, Kasur, Rawalpindi, Multan, Faisalabad, Bahawalpur, Gujranwala, Sargodha, and Karachi.
As the Safe City project grows, it is important to place it in the broader context of global smart city and smart policing initiatives and critically examine its proposed benefits and limitations. As a part of this analysis, it is crucial to identify the project, including its technological components and surveillance mechanisms, as an example of public infrastructure. Safe City initiatives are funded through public funds and are thus subject to public accountability, and deserve special scrutiny as they present potential risks and harms to citizens.
Tracing the Growth of Safe City Initiatives in Pakistan
We can trace the lineage of Huawei’s Safe City initiative through the emergence of technocratic “smart city” visions in the context of a long history of fears about urban safety and modern global city competitiveness. Huawei successfully marketed its Safe City as “a super-large comprehensive information-based management system established by the government’s safety agency to maintain harmony and security in society”. Over the past decade, it has established over 70 projects in over 50 countries, mostly in the Global South, where demonstrating safety is especially crucial for attracting foreign investment.
The first Safe City project in Pakistan was similarly conceived in the context of terrorism, specifically the aftermath of the Marriott Hotel bombing in 2008 in Islamabad. After a significant loan from the Chinese government, it lapsed for a few years until it was revived and finally completed in Islamabad at a cost of USD $125m by the National Database and Registration Authority (NADRA) and Huawei in 2016.
The Punjab Safe Cities Authority was also established in the same time frame, with an ordinance passed in 2015 and the completion of Lahore’s PPIC3 in January 2018. The PPIC3 was developed by Huawei, who won a contract for Rs. 12 billion (approximately USD $100 million), with consulting and technical support provided by National Engineering Services Pakistan (NESPAK) and Arup – a global architecture, engineering and design firm based in the UK. The Safe City projects in Islamabad and Lahore have inspired over a dozen similar projects across the country, most of them in Punjab. While only a few of the planned projects have been realised, these plans reveal a proposed urban surveillance infrastructure that would cover the majority of the country’s urban areas, with over tens of thousands of cameras installed in public spaces.
The Safe City projects promise to deter crime before it takes place, catch criminals, provide improved traffic management, and improve emergency service coordination and delivery through a number of technological components, including surveillance systems with facial recognition, dedicated telecommunication networks, data centers, drones, mobile applications, and intelligent transportation systems.
This technology assemblage, in turn, enables a number of operational and cultural transformations, such as better access to information, improved coordination, and ultimately, greater collaboration and accountability.
However, these purported benefits are largely aspirational at present, and raise serious questions about the tradeoffs being made in the name of security.
How safe are the Safe City initiatives?
There are three sets of issues facing Safe City initiatives in Pakistan: they threaten the human rights of those who live in Pakistan, introduce new risks and threats, and perhaps most worryingly, there is inadequate evidence that they have actually improved urban safety.
Threatening human rights
One of the fundamental issues with the Safe Cities initiatives, as with all mass surveillance programs, is that they threaten the basic human rights of privacy and freedom of expression, which are both constitutional rights in Pakistan as well, according to S.14(1) and S.19 of the Constitution. Pakistan’s Safe City initiatives already possess facial recognition technology, recognised as one of the most invasive forms of biometric identification, and provide inadequate data protection. In addition, the spectre of constant surveillance in public spaces and on online platforms threatens the freedom of expression and actually diminishes the feeling of safety.
There have been a number of high profile cases of data breaches from NADRA, which controls the Islamabad Safe City, as well as from PSCA. Questions have been raised about the security of this data for years, and there is still little in the form of personal data protection in Pakistan. The proposed Personal Data Protection Bill, 2020 is still under review, and even in its current form, would be inadequate to address the risks to individual and collective privacy from mass surveillance infrastructure. In addition, the Prevention of Electronic Crimes Act, 2016, also fails to protect citizens from surveillance and violates privacy.
The PSCA originally relied on a set of Electronic Data Regulations in 2016, which had a number of gaps. In 2020, it produced a set of Data & Privacy Protection Procedures (DP3), which seek to protect privacy rights and provide a number of data use, sharing, and handling guidelines. However, these guidelines are still inadequate in the absence of national data protection regulation. Moreover, Safe City initiatives and indeed all law enforcement agencies lack a data governance framework that describes where and how data would be stored by Safe City authorities, and how decisions about data will be made.
Types of data being collected by the Safe City:
- Personal data
- Vehicle data
- Traffic data
- Criminal profiles
- Parking data
Introducing new risks
Beyond the existing risks that surveillant assemblages present, they also introduce new risks to Pakistanis and all those who live in Pakistan.
The first risk is of cybersecurity threats that are inherent to all digital infrastructure, which are especially critical for public infrastructure like Safe City technologies that have access to critical, highly sensitive information with national security implications. The notion of an attack surface may be helpful here, which highlights the total number of potential entry points or vulnerabilities in a given technological system, whether in the form of hardware or software. The sheer number of objects associated with different Safe City projects – tens of thousands of cameras, microphones, RFID tags, mobile units, and other devices – as well as the interconnected network creates a high risk that needs to be addressed.
The second risk is of digital sovereignty. Most of Pakistan’s Safe City initiatives rely on technology provided by Huawei; this reliance effectively transfers significant power to a foreign entity. While Pakistan is not currently in a position to indigenously develop the technologies needed to run a surveillance infrastructure, it may need to provide safeguards to prevent against the existing systems being taken down. As noted later in the article, thousands of cameras in Lahore were taken offline because of a dispute with the vendor; this could be dangerous in the future.
Do Safe Cities work?
A lack of data
Beyond the fanfare, there is limited, low-quality data about what these Safe City initiatives have actually accomplished. This is surprising for an initiative that relies so heavily on data and technology.
Only one of the initiatives, PSCA, reports shares limited data through press releases that are published nearly verbatim by different news outlets. Despite PSCA’s claims, the government’s own data shows that crime is not significantly impacted. The number of rapes, abductions, murders, burglaries, and car snatchings has gone down, but other serious and violent crimes have increased. This points to a persistent issue of exaggerated claims with Huawei’s Safe City initiatives.
The data is also reported at irregular intervals and, in the absence of benchmarks, loses much of its potential to generate meaningful analysis and improve decision making. PSCA’s data sharing also represents a significant missed opportunity for open data and open government.
To take one of the numbers regularly provided as an example, PSCA issued 1.98 million e-challans to traffic violators in 2020, resulting in Rs. 192.73 million paid to the government. This represents an average of a Rs. 100 challan paid. Unfortunately, this reporting does not tell us whether this represents a good or bad outcome, even though the data that could and should be collected using PSCA’s state-of-the-art technologies can reveal a great deal. For example:
- How many traffic violations took place?
- Where did they take place?
- How many of the traffic violations were caught?
- How many of the e-challans were paid?
- What were the kinds of traffic violations that took place?
- How is PSCA data used in conjunction with existing research on road safety and traffic violations, such as this study on road traffic accidents in Lahore?
- What is the average number of traffic violations that takes place in a given city, neighborhood, or street in a year?
- What is the average number of e-challans that are issued annually? (Given that PSCA has been operational for at least 3 years, there should be enough data to provide a preliminary estimate).
In addition, it’s not clear that operational data, some of which would necessarily be confidential, is actually being collected and used to improve the work. For example, the locations and maintenance status of cameras and other sensors may be confidential, but it may be helpful to maintain records of uptime and service levels to build trust. Lahore’s Safe City project faced criticism last year because nearly half of the cameras were no longer functional; notably, this was due to a dispute with Huawei, which points to the risk of sovereignty identified above. There is also no information on how, if at all, the PSCA is affecting relationships between citizens and the police.
Deeper conceptual challenges
Beyond a lack of data, there are two deeper questions about Safe City initiatives.
The first is that they reflect a top-down, technocratic approach that relies uncritically on data and technology, even in the face of ample evidence that data is often unreliable or incorrect, can be misinterpreted, can reflect historical biases, and technologies can perpetuate harms. There is a long record of literature on the role of data and technology in policing that questions the use of new tools, especially surveillance, as a means of improving public safety.
While data and technology can be effectively used to help address the complex social challenge of urban insecurity, they do not represent long-term, systemic fixes to such insecurity because they do not address its underlying factors.
Indeed, such technological assemblages can be quite toothless without appropriate authorisation, as evidenced by the inability of the Islamabad Safe City to address mob violence in March 2016 when supporters of Mumtaz Qadri marched in the capital.
The second issue is that the Safe City initiatives have a limited perspective of safety and security, and indeed only address certain types of safety and security challenges in urban public spaces. The design of the Safe City initiative reveals a mode of thinking about violence as something that happens on the streets, usually from strangers; as if violence is primarily what bad people do. It doesn’t recognise the violence experienced by women who are subject to sexual harassment, for instance, nor does it account for state violence such as evictions. It focuses primarily on “spectacular violence”, such as terrorism or robbery, while largely ignoring “everyday violence”, such as domestic violence.
Public safety rests on addressing the fear of violence in everyday life, which is linked to people’s vulnerabilities, which in turn are linked to a lack of infrastructure and poor basic services. As a 2016 study titled “Gender and Violence in Urban Pakistan” notes , vulnerability is chronic, not accidental or extraordinary; and it is often created by intentional denial of citizenship, identity, basic services and human rights, as is experienced by Burmese and Afghan communities in Karachi and Islamabad. Religious minorities in Pakistan frequently face threats of religious violence; women are vulnerable due to poor mobility options, and their options are accordingly limited. As a study by the Digital Rights Foundation has described, the PSCA Public Safety App does not protect many of the women that are most vulnerable to violence on the street.
Indeed, the lack of access to services is a form ofviolence as well as a driver ofviolence, through competition for resources and associated mental health impacts which can be categorised as psychological violence.
Pakistan’s Safe City initiatives represent real digital transformation efforts in policing, but they are largely techno-solutionist, technocratic, top-down interventions that require more evidence and accountability. In addition, they must be placed in the context of the security state, where citizens are already under threat and are constantly surveilled by the government. In the absence of a digital rights framework, these surveillant infrastructures may threaten the human rights of Pakistanis and all those who live in Pakistan.
While Safe Cities are marketed as a huge advance to citizens, the real costs and benefits are not clear. Their emergence, along with much of the smart city vision, is linked to global city competition and the emphasis on investments that will attract foreign investors, businesses, or tourists rather than cater to the needs of existing residents.
This raises important questions about whose needs are prioritised, and whose safety matters – indeed, who is seen as “safe” and who is seen as a “threat” by the camera, and how this will translate into justice and security for all. A securitised public sphere defeats the notion of safe public spaces, and misses the point that safety comes not through aggressive policing and state surveillance, but through equitable social infrastructure and community trust.