Discussion Prompt: Should we ban the use of automated 

See all contributions to this question.

Modern video surveillance is a far cry from its clumsy predecessor. As technology has improved, and camera prices plummeted, surveillance en masse is likely coming to a city near you. Belgrade is one such city, experiencing the roll-out of thousands of cameras as part of a so-called “Safe Society” project. Installed without any public debate, nor a strong legal framework protecting digital and civil rights, concern by local civil society is high. Facial recognition in public spaces is one tool to fight crime, but residents must ask themselves, should this highly intrusive measure trump the privacy of all citizens?

Mass biometric surveillance can adversely affect a society — especially if in the case of Serbia, it is one with an already weak democratic tradition — and can cause serious violations of human rights. This is why it should be banned.

The digital age has brought about the idea that technology can exclusively be a force for good, helping us achieve nearly crime-free societies and peace among nations. Whether it’s preventing terrorist attacks or combating organised crime, one answer has been more surveillance of communications and the movements of citizens – most of whom are law-abiding. Now, entering the 2020s, the next targets of surveillance are our faces

Thousands of new eyes for Belgrade

Citizens of the Serbian capital Belgrade learned in early 2019 that their city will be covered with a thousand cutting-edge surveillance cameras in the following two years as part of the so-called “Safe Society” project. The project was unveiled without any prior public debate. What especially caught the public’s attention was the fact that these cameras — supplied by Chinese tech giant Huawei — will have facial and vehicle license plate recognition capabilities. The news was declared by high-ranking officials in internal affairs, the Police Director of Serbia and the Minister of Interior. The latter is one of the key figures of the ruling party and a close associate of President Vučić, which gave the announcement particular ‘weight’ in public. Since then, a citizen initiative known as “Thousands of Cameras” (“Hiljade kamera”), led by SHARE Foundation — a non-profit organisation dedicated to protecting digital rights, which I work for — has been actively challenging this surveillance system and demanding that such an intrusive technology be discussed in an open and inclusive setting before it is introduced.

Serbia does not have a long democratic tradition and features a problematic human rights record to this day. As a remnant of socialist Yugoslavia, which prioritised safety and security, privacy awareness is very low for most of the population. The country’s recent democratic backslide is also quite alarming. Earlier this year, Freedom House downgraded Serbia to a Transitional/Hybrid regime for the first time since 2003. On the World Press Freedom Index for 2020, Serbia is ranked 93rd out of 180 countries – another 3 places down from the previous year. In its report, Reporters Without Borders states that “Serbia has become a country where it is often dangerous to be a journalist”. Data protection and privacy do not rest on a long political tradition. A data protection law has existed in Serbia for just over a decade and the Commissioner for Information of Public Importance and Personal Data Protection (the national data protection authority) was established only 16 years ago, at first as a freedom of information complaints body. Furthermore, video surveillance in Serbia — a country currently negotiating EU membership — has seen its fair share of controversy, with camera footage often being abused, leaked in the pro-government media, or the cameras ‘conveniently’ not working at key moments (i.e. when powerful individuals could have been compromised by the footage). 

As the city administration’s infrastructure strategy is quite unpopular, the citizens of Belgrade, despite a general privacy lethargy, have paid attention to the new surveillance system; citizens and “Thousands of Cameras” activists have mapped hundreds of locations across Belgrade where cameras have been installed. This form of crowdsourcing provides more information about surveilled locations than the police itself has published. Photos of cameras from various Belgrade neighbourhoods are regularly posted on the “Thousands of Cameras” Twitter feed. With this alarming spread of cameras, we have to ask what the deeper implications of such surveillance are? And can it cause or cement irreversible changes in a world of declining democratic values, particularly in a country like Serbia?

The legal framework

Serbia has modernised its data protection legal framework by adopting the new Law on Personal Data Protection (LPDP) in late 2018; its full application began in August 2019. Drafted from a confusing mix of translated GDPR regulations and the EU Law Enforcement Directive, the text of Serbia’s new LPDP was controversial from the start, but at least it provided a more modern approach to data protection. Among its main flaws however, is the fact that the new LPDP does not specifically regulate two important aspects of data processing: biometric data and video-surveillance. Despite this, the law’s general provisions and principles should still apply to any data processing, such as a massive video surveillance system across Belgrade.

Before deploying a public space surveillance system, the LPDP obliges the data controller to conduct a Data Protection Impact Assessment (DPIA) and ask for the Commissioner’s opinion. However the Ministry of Interior of Serbia, which is the implementing body for the “Safe Society” project, failed to comply with the LPDP. It issued two DPIAs, both of which did not satisfy the Commissioner, who refused to approve. Despite this, the cameras were installed anyway. 

The latest information gathered from the second DPIA of the Ministry suggests that there will be more than 8000 different cameras and other devices in use, such as body cams, mobile cameras and vehicle-mounted cameras. Although facial recognition, i.e. automated detection of people’s faces from a video feed, is not yet rolled out by the Ministry, this feature is expected to be implemented by the end of the project. While little is known about the project’s timeline, this can be expected to be in the next two years.

Apart from the data protection issues related to facial recognition, we also need to ask whether these technologies are necessary and proportionate, particularly from the perspective of the European human rights framework and its underlying values. Is facial recognition in public spaces the only available measure that can be used to prevent serious crime and protect citizens? Can this highly intrusive measure trump the privacy of all citizens, effectively turning whole cities into mass surveillance zones?

All in all, the Belgrade surveillance camera system is of dubious legality, to say the least, because: 

  1. the actual purpose of introducing this system has not been clearly defined; 
  2. it has not been confirmed that the use of this system is necessary for the operations of state bodies; 
  3. its positive influence on the reduction of criminal offences has been overrated and its use is not proportionate to the risks related to the rights and freedom of citizens; 
  4. there is no law to begin with that defines that the police have the right to use smart surveillance in public places; and 
  5. the Data Protection Impact Assessment (DPIA) of the Ministry of Interior does not meet formal and material conditions defined by the law and was not approved by the Commissioner. 

Point of no return for human rights

Automated biometric video-surveillance may be the pinnacle of today’s techno-solutionism – trying to solve deep and complex social problems with often non-critical use of technology. Sadly, decision makers are usually blind to issues of ethical and legal nature, and far-reaching consequences, if a society sets public safety as its ultimate value. It is all the more troubling if they believe it can be achieved with technology such as mass video-surveillance. Once governments get a hold of such powerful technology, it might be impossible to completely remove it from their arsenal, even after successful legal challenges. In that regard, we can draw a parallel to blanket communications metadata retention — a highly controversial measure in terms of proportionality which is still alive and kicking in the EU, despite two CJEU judgements against it.1 Not to mention lucrative infrastructure deals required to install a massive video surveillance network, possibly in every larger city. This is particularly worrisome for countries such as Serbia, which are currently experiencing democratic backslides.

In addition to privacy, other associated human rights and freedoms, such as freedom of expression and the rights to protest and peaceful gathering, would also be affected in areas covered with automated video-surveillance. Imagine a scenario where a government keeps a biometric database of all citizens who attended anti-government protests; the ways in which this could affect their work, families, social relationships, and other aspects of everyday life are vast. Facial recognition also discriminates against minorities, further entrenching bias against disadvantaged communities and making them more vulnerable.

With its high risks and numerous adverse effects, especially once reaching a “point-of-no-return”, automated biometric video surveillance does not uphold the values of respect for human rights, equality and social justice of the EU and the Council of Europe. Therefore, banning automated video surveillance is the right step forward, especially when we take into account other worrying trends, such as the total surveillance of urban public spaces for law enforcement purposes, robots, drones and predictive policing.

The case of Belgrade and its thousands of cameras may prove to be only the beginning of a much wider trend, given the social tensions not just in Serbia, but in other countries as well. Freedom and democracy are on the decline, and our faces may be the last piece of the puzzle for establishing a dystopian society with unchecked surveillance and social control. At least, if we fail to resist such a system before it is too late.

[1] Notwithstanding this week’s judgment in the cases of La Quadrature du Net, the French Data Network, Ordre des barreaux francophones et germanophone, and Privacy International.