Political freedom, generally speaking, means the right to be a participator in government, or it means nothing.Footnote 1
7.1 Introduction
In 2018, police in India reported that the roll out of facial recognition technology (FRT) across New Delhi enabled their identification of 3,000 missing children in just four days.Footnote 2 In the United Kingdom, South Wales Police used live FRT to scan over 50,000 faces at various mass gatherings between January and August 2019 and identified nine individuals for arrest.Footnote 3 The Chinese Sharp Eyes programme, ‘omnipresent, fully networked, always working and fully controllable’, can take less than seven minutes to identify and facilitate apprehension of an individual among a population of nearly 5 million people.Footnote 4 In Moscow, 105,000 FRT-enabled cameras have monitored and enforced COVID-19 self-isolation orders,Footnote 5 with at least 200 violators being identified.Footnote 6
As protest movements are gaining momentum across the world, with Extinction Rebellion, Black Lives Matter, and strong pro-democracy protests in Chile and Hong Kong taking centre stage, many governments – both in the West and in the East – have significantly increased surveillance capacity of the public sphere. City streets and squares, stations, and airports across the globe, and social media and online platforms have become equipped with sophisticated surveillance tools, enabled and made legal through a myriad of complex and ever-expanding ‘emergency’ laws. Irrespective of whether these events and/or political strategies are framed as ‘emergencies’ such as the ‘war on terror’ with its invisible geopolitical enemies for 9/11, or whether they were pro-democracy or anti-racism protests or connected with COVID-19, the state resort to technology and increased surveillance as a tool to control the masses and population has been similar. Examples from varied countries – ranging from China, Russia, and India to the United States and the United Kingdom – tell us that recent technological advances have enabled authoritarian and democratic governments alike to build omnipresent biometric infrastructures that systematically monitor, surveil, predict, and regulate the behaviour of individual citizens, groups, or even entire populations. In this chapter, I focus on the chilling effect of FRT use in public spaces on the right to peaceful assembly and political protest. While technological tools have transformed protest movements widely, both amplifying and undermining them,Footnote 7 in this chapter I only focus how protest movements have been tackled with FRT and my emphasis is on political protests and public spaces. Pointing to the absence of oversight and accountability mechanisms on government use of FRT, the chapter demonstrates how FRT has significantly strengthened state power. It draws attention to the crucial role of tech companies in assisting governments in public space surveillance and curtailing protests. I argue for hard human rights obligations to bind these companies and governments, to ensure that political movements and protests can flourish in the post-COVID-19 world.
7.2 Undermining Protest Movements with FRTs
Live automated FRT, rolled out in public spaces and cities across the world, is transforming modern policing in liberal democracies and authoritarian regimes alike. The technology augments traditional surveillance methods by detecting and comparing a person’s eyes, nose, mouth, skin textures, and shadows to identify individuals.Footnote 8 The live automated facial recognition can instantaneously assess the facial biometric data in the captured images against a pre-existing ‘watchlist’ and flag it to police officers. Some FRT tools go further, purporting to classify people by gender or race or make predictions about their sexual orientation, emotions, and intent.
This FRT has been used to tackle protest movements globally. For example, the US company Geofeedia has been marketed to law enforcement ‘as a tool to monitor activists and protestors’,Footnote 9 incorporating FRT use with Twitter, Facebook, and Instagram databases.Footnote 10 Rasheed Shabazz, an activist and journalist, believes that his arrest near the Black Lives Matter protests in Oakland in 2014 was as a result of the Geofeedia software.Footnote 11 This same software was also used to monitor civil unrest after the police killing of Freddie Grey and link protesters with their social media profiles.Footnote 12 Similarly, in 2020, during the protests following the killing of George Floyd in Minneapolis, Minnesota, several people were arrested and charged after being identified through the use of FRT.Footnote 13 In another case, the Detroit Police Department used FRT to identify a Black Lives Matter protester who was arrested and charged with reckless driving and resisting arrest.
Similarly, FRT has been used in many other countries. For example, ‘habitual protesters’ in India are included in a dataset used to monitor large crowds,Footnote 14 which is composed of ‘miscreants who could raise slogans and banners’.Footnote 15 This database was used to identify dissidents at a prime ministerial rally in December 2019,Footnote 16 and also resulted in the detention of a ‘handful’ of individuals charged with violent crimes when it surveyed protests in New Delhi and Uttar Pradesh.Footnote 17 The Hong Kong police used FRT cameras to identify protesters and track their movements during the 2019 pro-democracy protests, which drew criticism from human rights advocates who argued that it violated the protesters’ right to privacy and could lead to their persecution.Footnote 18 In 2019–2020, FRT cameras were also used in Chile to monitor and identify protesters participating in demonstrations and civil unrest, known as the Estallido Social.Footnote 19 The cameras were installed in public areas, including train stations and street corners, by the Chilean government to track individuals who were suspected of participating in protests or other forms of civil disobedience. In the face of mounting criticism and protests against the use of this technology, the Chilean government announced that it would suspend the use of facial recognition cameras in public spaces in early 2020.
In all these cases, FRT allowed the authorities to quickly identify individuals who were wanted for questioning or arrest. The cameras were linked to a central database containing photos and personal information of individuals who were known to have participated in previous protests or other activities that the government deemed to be illegal. Such use of FRT cameras sparked controversy and concern among civil liberties groups and privacy advocates, who argued that the technology was being used to stifle dissent and violate the rights of protesters to peacefully assemble and express their opinions. Despite these concerns, governments typically defend FRT use by framing it as a necessary measure to maintain ‘public safety’ and order during a time of civil unrest.
In addition to such ‘top-down’ surveillance by public authorities in USA, India, Hong Kong, and Chile, ‘horizontal’ modes of surveillance have become increasingly popular.Footnote 20 This involves partially outsourcing surveillance functions to individuals and/or tech companies. A vivid example of such outsourced surveillance was the 2020 Black Lives Matter protests in Dallas, during which the police department asked individuals on Twitter to send them videos from protests that showed ‘illegal activity’.Footnote 21 A larger-scale example was seen in the aftermath of the 2010 Canadian Winter Olympics riots, in which closed-circuit television (CCTV) footage was used to identify offenders, and private individuals sent the Vancouver Police Department thousands of images and helped them scour social media.Footnote 22 Similarly, tech companies such as Facebook, Twitter, and Instagram have been crucial in surveillance of protesters, as the widespread use of social media has made the monitoring of protest and dissident activities significantly easier.Footnote 23 For example, in 2014 and 2016, the US government obtained two patents that may facilitate its ability to use social-media to predict when a protest will break out.Footnote 24
Protest movements in USA, Hong Kong, Chile, and beyond have also operated in the shadow of the global COVID-19 pandemic, and together raised questions about unprecedented levels of government power and the expanding regime of mass surveillance in public spaces. The COVID-19 pandemic has given governments a further impetus to explore FRT’s health-related uses – from monitoring compliance with quarantine or social-distancing requirements to tracking (in conjunction with other biometric technologies such as thermal scanning) those who are potentially infected. COVID-19 and the latest protests in Hong Kong, Chile, and the United States have redefined the boundaries of mass surveillance and biometric tracking globally, with irreversible implications for the future exercise of government power and surveillance.
7.3 Lack of Regulation and Dangers of FRT
Despite the increasing deployment of FRT in many city squares and streets across the globe, as many chapters in this book demonstrate, FRT use is not yet regulated. Law enforcement agencies around the world are experimenting with FRT with discretion and on an ad hoc basis, without appropriate legal frameworks to govern its use nor sufficient oversight or public awareness.Footnote 25 For example, there are currently no federal regulations in the United States governing the use of FRT by law enforcement.Footnote 26 In March 2019, two US senators introduced the Commercial Facial Recognition Privacy Act, intended to ban developers and providers of commercial FRT from collecting and sharing data for identifying or tracking consumers without their consent.Footnote 27 However, this only focussed on the commercial use of FRT. Similarly, in the EU, regulation of FRT has been very limited. In February 2020, a draft EU White Paper on Artificial Intelligence appeared to call for a discussion about a temporary five-year ban on facial recognition. However, the final draft of this paper removed mention of such a moratorium.Footnote 28
This lack of oversight of FRT use by public bodies can lead to abuses of power and violations of fundamental rights and civil liberties. As many chapters in this book demonstrate, FRT use can result in discriminatory treatment and undermining of privacy and due process, as well as other concerns. Indeed, the dangers of FRT are gradually being recognised by courts. For example, law enforcement’s use of automated FRT was successfully challenged in 2020 in R (on the application of Bridges) v. Chief Constable of South Wales Police ([2020] EWCA Civ 1058) (‘Bridges’) case, where the Court of Appeal held that the use of automated FRT by South Wales Police was unlawful because it was not ‘in accordance with law’ for the purposes of Article 8 of the European Convention on Human Rights.Footnote 29 In addition, South Wales Police had failed to carry out a proper Data Protection Impact Assessment and had not complied with the public sector equality duty.Footnote 30 While Bridges is the first successful legal challenge to police use of automated FRT worldwide, fresh lawsuits brought by non-governmental organisations in the United States and France are still pending, and they might provide different judicial responses to regulation of police FRT use.Footnote 31
Some jurisdictions have already regulated and limited FRT use by law enforcement. In the United States, for example, the cities of San Francisco and Berkeley have banned local agencies (including transport authorities and law enforcement) from using FRT,Footnote 32 some municipalities in Massachusetts have banned government use of facial recognition data in their communities,Footnote 33 and other US states (California, New Hampshire, and Oregon) have instituted bans on facial-recognition technology used in conjunction with police body cameras.Footnote 34 The United Kingdom also has an Automated Facial Recognition Technology (Moratorium and Review) Bill,Footnote 35 proposing to ban the use of technologies. Yet its future remains uncertain.
Therefore, not only civil right advocates, but also the courts and politicians widely recognise that FRT can be easily misused by law enforcement to target certain groups of people, such as political activists or marginalised communities, and such targeting often leads to further discrimination and injustice. Importantly, the growing prevalence of surveillance through FRT has a chilling effect on public discourse by threatening the right to protest anonymously; a notion fundamental to protest movements.
7.4 Protest Movements, Public Space, and the Importance of Anonymity
Protest movements are collective actions undertaken by a group of people who come together to express their dissent, raise awareness, and advocate for change around a particular issue or cause.Footnote 36 These movements can take many different forms, ranging from peaceful demonstrations, marches, and rallies to civil disobedience, strikes, and other forms of non-violent resistance. Protest movements can emerge in response to a wide range of social, economic, political, and environmental issues. Some of the most common causes of protest movements include discrimination, injustice, corruption, inequality, environmental degradation, and war. Contemporary examples are Occupy Wall Street (2011), Arab Spring (began in 2010), Black Lives Matter (began in 2013), and the Hong Kong pro-democracy movement (began in 2019). Protest movements can also be motivated by a desire to promote social change, challenge existing power structures, and hold those in authority accountable for their actions.
Throughout history, protest movements have played a critical role in advancing social progress and promoting human rights. They have helped to raise awareness of important issues, mobilise public opinion, and influence policy and legislative changes. Examples of protest movements from history include the civil rights movement of the 1950s–1960s, the women’s suffrage movement of the late nineteenth and early twentieth centuries and Vietnam anti-war protests (1960s). Today, protest movements continue to be an important tool for promoting social change and advocating for a more just and equitable world.
Protests movements require a tangible and accessible location, typically in the streets and other public places. Public space has always been central to social movements and political protests as a practical place for citizens to gather and as a symbolic place connected to wider democratic values. It provides a physical location where individuals can come together to voice their dissent, express their grievances, and demand change.Footnote 37 By occupying public spaces, protesters can create a visible and disruptive presence that draws attention to their cause, and can also serve as a symbolic representation of their struggle.
Public spaces, such as city squares, parks, and streets, are often central to the social and cultural life of a community, and their use for protests can be a powerful statement of the collective will of a group of people. Thus, public spaces are the ‘ultimate area of societal interaction’ and occupy a symbolic place in society given their accessibility, openness, and, according to Jens Kremer, inherent freedom.Footnote 38 When protesters occupy public spaces, they are asserting their right to participate in the democratic process and to be heard by those in power. In interrupting these public spaces, protesters ‘touch upon the very core of the current structure and organization of social systems, namely the balance of power, rule of law and democratic governance’.Footnote 39 It questions the ability of government authorities to maintain the integrity of these shared spaces,Footnote 40 thus challenging existing power structures.
Historically, protesters have taken the right to protest anonymously largely for granted; a right that is now becoming increasingly more fragile. The right to anonymity has been fundamental to social movements and protesting, as these events require the population to feel confident and safe in their ability to gather in public spaces and manifest their disagreement with the status quo. This is impossible if they fear surveillance tools can be weaponised against them to suppress and punish their dissent. The sense of safety necessary to facilitate robust democratic participation stems from an understanding that an individual, in the act of demonstrating, is expressing something larger than themselves by joining in a collective. They thus sacrifice their individual voice for the benefit of social disruption, and in return are granted the key right that protesters have enjoyed for centuries; the right of anonymity. The anonymity earned by protesters in public spaces has been increasingly challenged and eroded by surveillance infrastructure.
While the relative anonymity of the individual during protest gatherings has typically ‘neutralised’ the effect of surveillance, they have been increasingly subject to ‘counter-neutralization technologies’ that require those individuals to take more active steps to circumvent identification.Footnote 41 Of course, protest movements have long devised resistance strategies against surveillance. For example, protesters can break a system a surveillance system by flooding it, rendering surveillance inoperable or impractical.Footnote 42 Typical examples include crude forms of neutralisation such as disabling phone lines, wearing masks, and destroying cameras. For example, Hong Kong protesters in 2019 used lasers and broke smart lampposts that they believed contained FRT software.Footnote 43 With FRT, protestors are given two choices: first, they can wear a mask and risk arrest and the collection of biometric information in the form of criminal records, or second, they can do without a mask and risk collection of biometric data through FRTs.Footnote 44
Surveillance technologies dealing with political protests have become the norm in many countries, and scholars have theorised about the chilling effect of surveillance on dissent.Footnote 45 Monitoring, tracking, and detaining individual protesters for their actions in public places significantly shifts the power balance between the state and individuals. Surveillance of political protests undermines the individual as a ‘free autonomous citizen’, and negatively impacts democracy and the rule of law.Footnote 46 Protestors become disempowered in relation to their body and biological information,Footnote 47 which is not only threatening, in one sense, to discrete individuals,Footnote 48 but in another to discretise protesters, breaking down their collective image. Pervasive surveillance tools can be understood as disciplinary, as they are able to threaten and realise retribution against individual protesters who would otherwise have been lost in a sea of voices, but it is in another sense indicative of a ‘controlled’ society in which surveillance is ubiquitous.Footnote 49
7.5 Protecting Protesters from Abuse: Potential Ways to Regulate
Given the danger that FRT surveillance in public spaces poses to political protests, the rights to peaceful assembly and association, and wider democratic participation, legislatures should regulate or entirely ban the use of FRT in policing and law enforcement. Regulation of FRT use is a necessary step to ensure the chilling effect of FRT on political expression and freedom of assembly is eliminated.
The chilling effect on freedom of speech and assembly is even stronger in some jurisdictions, such as Australia. This is because, unlike many other jurisdictions discussed in this book, Australia has no human rights protection enshrined in its Constitution and no national human rights legislation.Footnote 50 Only three out of eight Australian states and territories have state-level human rights Acts. For this reason, in its recent report, the Australian Human Rights Commission has urged Australia’s federal, state, and territory governments to enact legislation regulating FRT.Footnote 51
What are the ways to protect protesters and protest movements from abuse by public authorities? Recent literature related to AI and accountability has recommended several avenues, including regulation,Footnote 52 the development of technical methods of explanation,Footnote 53 the promoting of auditing mechanisms,Footnote 54 and the creation of standards of algorithmic accountability in public and private bodies.Footnote 55 Law, of course, should also play a role.
7.5.1 Privacy Law
Privacy law provides one avenue to regulate police use of FRT in public spaces.
Scholars have long argued that public activities deserve privacy protections, and that the simple act of being ‘in public’ does not negate an individual’s expectation of privacy.Footnote 56 However, as Jake Goldenfein in Chapter 5 of this book suggests, privacy law has severe limitations when regulating the use of FRT. For example, the US Fourth Amendment, under current Supreme Court jurisprudence, has been viewed as an unlikely protection against FRT for two reasons: first, jurisprudence has typically ignored pre-investigatory surveillance,Footnote 57 and secondly, it has failed to encompass identification of information already exposed to the public.Footnote 58 In relation to the Fourth Amendment, Douglas Fretty questions whether Constitutional protection will require the Supreme Court to confirm the ‘right of the people to be secure’ or simply display how insufficient the Fourth Amendment is in safeguarding individuals beyond the scope of their private spaces.Footnote 59 Drawing on recent US Supreme Court cases concerning GPS tracking and other technologies and the Fourth Amendment,Footnote 60 Andrew Ferguson suggests that the Supreme Court is cognisant of the need to adapt Fourth Amendment jurisprudence to emerging technologies.Footnote 61 He identifies six key principles in adapting the Fourth Amendment to deal with modern concerns. First, technological searches cannot be viewed as equivalent to pre-technological police investigatory modes.Footnote 62 Secondly, there must be a general presumption against the large-scale aggregation of data.Footnote 63 Thirdly, there must be a general presumption against the long term storage and ongoing use of aggregated data.Footnote 64 Fourthly, the ability to track and trace an individual must be a relevant factor in considering the application of the Fourth Amendment.Footnote 65 Fifthly, the concept of anti-arbitrariness must be transposed to a digital setting to act against automated technologies that do not require probable cause.Footnote 66 Sixthly, monitoring technologies must not be so over-reaching as to grossly permeate civil society.Footnote 67 However, the Fourth Amendment offers limited support in the protection of protest movements from FRT surveillance in public spaces.
7.5.2 Discrimination Law
Could discrimination law provide a better avenue to regulate police use of FRT in public spaces? The emerging consensus in an increasing body of academic research is that FRTs are not ‘neutral’,Footnote 68 but instead reinforce historical inequalities.Footnote 69 For example, studies have shown that FRT performs poorly in relation to women, children, and individuals with darker skin tones.Footnote 70
This bias and discrimination can be introduced into the FRT software in three technical ways: first, through the machine learning process, based on the training data set and system design; secondly, through technical bias incidental to the simplification necessary to translate reality into code; and thirdly, through emergent bias that arises from users’ interaction with specific populations.Footnote 71 Because the training data for FRTs in the law enforcement context comes from photos relating to past criminal activity,Footnote 72 minority groups and people of colour are over-represented in FRT training systems.Footnote 73 In some jurisdictions, such as the United States, people of colour are at a much higher risk of being pulled over,Footnote 74 searched,Footnote 75 arrested,Footnote 76 incarcerated,Footnote 77 and wrongfully convicted than white women.Footnote 78 Therefore, police use of FRT to repress political protests can produce a large number of false positives – as it is already functioning in a highly discriminatory environment; and this can impact on the freedom of assembly and association of those already marginalised and discriminated against. However, discrimination law alone offers limited support in the protection of protest movements from FRT surveillance in public spaces.
7.5.3 Holding Private Companies Accountable for FRT Surveillance of Public Spaces
Private actors are also playing a role in the increasing surveillance of public spaces by stifling protest movements and political participation worldwide, and we need to insist on holding them accountable. Private companies, such as telecommunications service providers and tech giants, have been co-operating with law enforcement agencies and developing the technical infrastructure needed for public space surveillance. This includes police purchasing and using privately developed FRT technology or image databases, both of which often happen in secret. For example, IBM, one of the world’s oldest (and largest) technology companies,Footnote 79 has recently collaborated with repressive governments by providing FRT software. Best known for its famous personal computers, in recent years the company’s focus has shifted to AI and FRT.Footnote 80 A detailed report by The Intercept published in March 2019 revealed that in 2012 IBM provided police forces in the Philippines with video surveillance technology, which was subsequently used to perpetuate President Duterte’s war on drugs through extra-judicial killings.Footnote 81 The brutal and excessive crime suppression tactics of the Davao police were well known to local and international human rights organisations.Footnote 82
At the time, IBM defended the deal with Philippines, saying it ‘was intended for legitimate public safety activities’,Footnote 83 but claimed that it had ceased provision of its technology to the Philippines in 2012. However, it took at least several years for IBM to stop providing general purpose FRT software to law enforcement (e.g., IBM mentioned its Face Capture technology in a public disclosure in 2013 and 2014, related to its Davao City project).Footnote 84 The company’s practice of providing authoritarian regimes with technological infrastructure is not new and dates back to the 1930s, when IBM supplied the Nazi Party with unique punch-card technology that was used to run the regime’s censuses and surveys to identify and target Jewish people.Footnote 85
Because of such close (and often secretive) collaboration between private tech companies and governments, we need to think of new ways to hold the companies providing the FRT infrastructure accountable – not just in aspirational language, but in law. Currently, in many countries, the application of human rights laws is limited to government bodies only (anti-discrimination and data protection laws being the primary exceptions of horizontal application).Footnote 86 The same is true of international human rights law. This leaves private companies in the human rights gap. However, as I have argued elsewhere in detail, existing efforts focussing on voluntary ‘social and corporate responsibility’ and ethical obligations of private tech companies are insufficient and incapable of tackling the challenges that these technologies pose to freedom of expression and association.Footnote 87 Moreover, a lot of those efforts have merely been ‘transparency washing’ – performatively promoting transparency and respect for human rights while acting in ways that undermine both.Footnote 88
The problem of the human rights gap is greater in some jurisdictions, such as Australia, which lacks a federal level human rights framework and where governments often remain unaccountable for public space surveillance. Therefore, we need to demand change and accountability from governments, police, and tech companies. We should not continue to rely on the ‘goodwill’ of tech companies, when they promise to ‘respect’ our right to protest and our freedom of association and assembly. We need to demand hard legal obligations for private actors because of the significant role they play in increasing public space surveillance and infrastructure. We need data protection and human rights laws that bind companies, to ensure that political movements and protests can flourish and that communities whose rights to peaceful assembly and association have been curtailed via FRT technologies can access an effective remedy.
7.5.4 Outright Bans on Police Use of FRT
Of course, even with all the limits that could be placed by law, police use of FRT in public spaces is problematic in itself – owing to the centrality of public space and anonymity for protest movements. It is thus not surprising that many scholars, activists, regulators, and politicians have turned to arguing for bans on FRT use. For example, US scholar Ferguson advocates for a blanket ban on facial surveillance, a probable cause requirement for facial identification, a ban or probable cause-plus standard for facial tracing, and limitations to facial verification at international borders in addition to increased accountability for error, bias, transparency, and fairness.Footnote 89
Proposals to ban FRT have also been coming from sources outside the academic realm; with informal resistance groups such as the developers of the website Fight for the Future running a project called Ban Facial Recognition, which operates an interactive map of where and how the government is using FRT around the United States.Footnote 90 Further, the United Kingdom’s Equality and Human Rights Commission,Footnote 91 and the Australian Human Right Commission,Footnote 92 have recently called on governments to introduce a moratorium on the use of FRT in policing and law enforcement before legislation regulating the use of FRT and other biometric technology is formally introduced.
7.6 Conclusion
If the government and law enforcement can resort to FRT without any restrictions or safeguards in place, the right to protest anonymously will be curtailed and political discourse in our democracies will be stifled. For example, the High Court of Australia – Australia’s apex court – has emphasised the centrality of the right to protest to Australian democracy: besides casting their vote in elections, Australians have no other avenues through which to voice their political views.Footnote 93 Adapting Hannah Arendt’s famous quote used at the beginning of this chapter, political freedom must enable a right to participate in government. And in many instances, the only way to do that, in addition to voting, is through political protest.
Before FRTs develop further and become even more invasive, it is imperative that this public surveillance infrastructure is limited. We need laws restraining the use of FRT in our public spaces, and we need hard legal obligations for those who develop and supply law enforcement with them. The reforms could start with an explicit ban (or at least suspension) on police use of FRT in public spaces, pending independent scrutiny of the discriminatory impacts the technology may have against women and other protected groups.Footnote 94 These proposed changes are not drastic. In fact, they are a modest first step in the long journey ahead to push back against escalating surveillance of the public sphere worldwide.