‘I started seeing shadows everywhere’: The diverse chilling effects of surveillance in Zimbabwe

Abstract

Recent years have witnessed growing ubiquity and potency of state surveillance measures with heightened implications for human rights and social justice. While impacts of surveillance are routinely framed through ‘privacy’ narratives, emphasising ‘chilling effects’ surfaces a more complex range of harms and rights implications for those who are, or believe they are, subjected to surveillance. Although first emphasised during the McCarthy era, surveillance ‘chilling effects’ remain under-researched, particularly in Africa. Drawing on rare interview data from participants subjected to state-sponsored surveillance in Zimbabwe, the paper reveals complex assemblages of state and non-state actors involved in diverse and expansive hybrid online–offline monitoring. While scholarship has recently emphasised the importance of large-scale digital mass surveillance, the Zimbabwean context reveals complex assemblages of ‘big data’, social media and other digital monitoring combining with more traditional human surveillance practices. Such inseparable online–offline imbrications compound the scale, scope and impact of surveillance and invite analyses as an integrated ensemble. The paper evidences how these surveillance activities exert chilling effects that vary in form, scope and intensity, and implicate rights essential to the development of personal identity and effective functioning of participatory democracy. Moreover, the data reveals impacts beyond the individual to the vicarious and collective. These include gendered dimensions, eroded interpersonal trust and the depleted ability of human rights defenders to organise and particulate in democratic processes. Overall, surveillance chilling effects exert a wide spectrum of outcomes which consequently interfere with enjoyment of multiple rights and hold both short- and long-term implications for democratic participation.

Introduction

Recent years have witnessed the growing ubiquity and potency of state surveillance measures. Much of this growth has been driven by the increasing sophistication and decreasing costs of advanced digital technology. At the same time that surveillance technologies have grown rapidly, their impacts and harms have been understood in relatively static terms, most commonly (albeit not exclusively) through the lens of ‘privacy’. Additionally, while much relevant scholarship has emphasised the tools and practices of surveillance, there remains markedly less emphasis on the experiences of those subjected to such practices.

This paper examines the under-researched issue of surveillance in the Zimbabwean context. It evidences and unpacks the chilling effects of such practices through the rarely heard voices of those affected by authoritarian surveillance measures. In doing so, the paper seeks to offer several interventions into how surveillance practices are analysed in the digital era. This is pertinent given the increased adoption of AI-enabled digital surveillance technologies, and the absence of associated research into their impacts. As surveillance does not typically occur in an exclusively digital context, it is hoped that a better understanding of the impacts of surveillance will elucidate issues likely to arise in the AI-enabled digital era. As digital surveillance becomes more pervasive, and state surveillance shifts from an individual (or small group) focus to a society-wide capability, the societal effects of surveillance discussed here are likely to be particularly relevant.

A principal focus of the paper concerns the empirical identification and analysis of the chilling effects of surveillance on the enjoyment and exercise of human rights essential to the development of personal identity and the effective functioning of participatory democracy; in particular, the right to freedom of expression, and the right to freedom of assembly. ¹ Focusing on the links between expression and assembly highlights that surveillance is not only a matter of individual privacy or individual expression, but something with a direct impact at a societal level, particularly on democratic processes. Renewed emphasis on ‘chilling effects’ then, is argued to surface a more complex range of harms and rights implications for those who are, or believe they are, subject to surveillance. As such, the paper intends that a more developed, and evidenced, understanding of the chilling effect, and the impact this exerts on human rights protections, will facilitate a better understanding of the consequences of surveillance while offering an initial evidential base to support future legal interventions and planning for the AI-era.

Second, Edward Snowden's 2013 revelations catalysed a heavy emphasis on online and other platform surveillance in the academic literature. The incursion of such practices into everyday life has been increasingly documented. Perhaps most commented on among these is Zuboff's (2019) exegesis of how online surveillance has become the organising principle of capital accumulation in the digital age. This paper recognises that such analyses of indiscriminate and comprehensive online monitoring are indeed vital to understanding the circuits of advanced capitalism, the reproduction of inequalities and the impact on individuals’ rights. However, it also argues that advanced corporate digital monitoring practices should not be regarded as an exemplar for all forms of contemporary surveillance. Emphasis on commercial surveillance practices may risk downplaying the enduring role of state surveillance, as it arguably does in Zuboff's work. This paper argues that the reality for many in the global south is that the cheapness of surveillance labour has invigorated traditional forms of human surveillance and these intersect with advanced online monitoring practices, resulting in the emergence of new human-digital surveillance ensembles. As a corollary, accounting for such socio-political contexts reveals the ‘reinstatement’ of the human – for example, through the infiltration of closed and encrypted platform communications – which may further undermine technology-focused solutions aimed towards protecting citizen's rights online.

The human rights situation in Zimbabwe is well known, and instances of torture, arbitrary detention and enforced disappearance are widely documented. Those targeted include journalists, political actors, human rights defenders and ‘ordinary’ citizens. By contrast, state-sponsored surveillance in Zimbabwe is an issue veiled in secrecy, leading to misconceptions about the extent of surveillance and the actors involved. While detailed research and literature on the role of surveillance in human rights violations is limited there has been growing attention brought to digitally-assisted state surveillance practices in the region by civil society actors (see below) and through the media. The latter, for example, has revealed how Chinese-backed infrastructure projects in the region have included the export of advanced digital surveillance tools that are increasingly integrated into state security apparatus (Parkinson et al., 2019). Through empirical evidence of first-hand accounts, this paper contributes detail on the nature of surveillance in Zimbabwe while recording manifest impacts of such practices – both actual and perceived – on individuals and society as a whole aiming to facilitate a more nuanced understanding of the chilling effects of surveillance, and articulate the implications for core democratic rights such as the right to freedom of expression, and the right to freedom of assembly.

The paper begins by setting out current understandings of the chilling effects of state surveillance practices. This focuses on the definitional and epistemological debate concerning the identification of the presence and impact of chilling effects, and makes the case for a greater focus on in-depth qualitative research which reveals the nuance of such effects and expands research into a greater breadth of contexts. The research method is then outlined before moving on to the findings.

Overall, the paper evidences the array of chilling effects experienced by those under surveillance in Zimbabwe. It argues that chilling effects should not be considered as holding a binary character – where they either exist or do not – but instead as a wide spectrum of outcomes which consequently interfere with the enjoyment of multiple rights and have both short- and longer-term implications for democratic participation.

What is the ‘chilling effect’?

The term ‘chilling effect’ has no universally agreed definition. It is used in this paper to explain how the fear (or possibility) that one is being watched affects an individual's conduct, impacting behaviours such as what they say, what websites they visit, what materials they post, what comments they make, who they interact with, and if, or how, they engage in political opposition. First appearing in McCarthy era US court cases, the term ‘chilling effect’ originated and remains most recognised in relation to adjudications of free expression. Chills can be both benign and invidious, with benign chills beneficially deterring unlawful acts such as hate speech or defamation (Townend, 2017). Invidious chills, however, can be understood more widely as a deterrent effect created by surveillance activities, which, through fear of persecution, punishment, or other harm, deters participation in activities – such as democratic life – which are typically protected or encouraged (Kendrick, 2013; Schauer, 1978). It is, however, important to emphasise that an immediate fear of harm or punishment is not a necessary prerequisite for chilling effects to occur. In certain contexts, it may simply be that individuals do not wish to gain the attention of state authorities and so modify their speech accordingly. As such, knowing or being suspicious that state surveillance exists is itself considered by some sufficient to create a demonstrable chilling effect (Kaminski and Witnov, 2015; Solove, 2007).

Within surveillance studies, self-censorship and modifications of individual behaviour following exposure to surveillance have been theoretically framed, most prominently through Foucault's metaphorical application of the Panopticon. For Foucault (1977), Panopticism characterises how a subject's awareness of possible surveillance invites self-regulation, self-discipline and behavioural modification. For some, this concept has attracted comparison to chilling effects given how the presence of surveillance alone, absent any material coercive or oppressive applications, may stimulate self-censorship among subjects (Manokha, 2018; Stoycheff et al., 2018). While Foucault's metaphor for pervasive surveillance is ubiquitous within academic commentary, it has drawn detractors. These have variously challenged its continued relevance (Haggerty, 2006) and considered it an ‘unhelpful shadow on analyses of surveillance and control’ (Fussey, 2013: 356). ²

Perhaps the most relevant critique of panopticism – particularly as applied to ‘mass’ surveillance – concerns failure to emphasise the uneven distribution of the surveillance gaze and its varying effects across social locations. It also assumes self-disciplinary effects, leaving questions regarding the extent to which subjects are ‘retrained’ and reintegrated towards docility or ill-defined norms rather than excluded (Bigo, 2006). Panopticism also downplays defiant adaptations and acts of resistance following revelations of state surveillance, such as those following the 2013 Snowden revelations (Horowitz, 2017). Notions of retraining and conformity to the norm conceal the nuanced, subtle modification of behaviour – rather than blanket silencing and conformity – that, for some, are characteristic of chilling effects (Kendrick, 2013; Stoycheff, 2016). As such, panopticism has become a somewhat anachronistic, and therefore inadequate, lens for considering the surveillance harms examined in this paper. Exploring the impact of surveillance through the wider reaching lens of ‘chilling effects’ therefore allows for accounts of behaviour change to emerge beyond the individual under surveillance, examining how the impacts resonate into the community, interfering with the collective exercise of rights. Despite efforts of surveillance scholars to demonstrate and document the social control and discrimination enabled and created by surveillance, others have argued that society still lacks a sufficient ‘account of when and why surveillance is problematic to help us see when we should regulate and when we should not’, and that what does exist fails ‘to speak in terms that are likely to influence the law’ (Richards, 2013: 1935). Utilising the lens of chilling effects responds specifically to this concern by enabling a direct examination of the impact of surveillance on human rights protections. As a result, it not only facilitates a wider, more nuanced understanding of the consequences of surveillance but opens up possibilities for potential legal responses.

Identifying and accounting for the chilling effects of surveillance

Despite the emergence of ‘the chilling effect doctrine’ within the US courts, there has been little success in cases which respond to state surveillance activities. Cases invoking chills to constitutional rights have been dismissed due to lack of standing (Penney, 2016), or because evidencing chills to fundamental rights has been problematic, with the results deemed ‘subjective’ (Laird v. Tatum, 1972) or ‘too speculative’ (Clapper v. Amnesty, 2013). Similarly, the European Court of Human Rights has yet to really engage with the chilling effects of large-scale surveillance – particularly at a societal level – and the effects of surveillance continue to be evaluated primarily through a right to privacy lens. Furthermore, as Penney (2016) points out, this ‘judicial scepticism’ (Kaminski and Witnov, 2015: 482) over the provability of chilling effects is not confined to the courts but also persists throughout legal scholarship (Solove, 2007). This has created a particular challenge for those seeking to ensure chilling effects-related harms are adequately reflected in litigation: how to prove something that did not happen would have otherwise occurred (e.g. participation at a protest), and that it did not happen because of a particular concern (e.g. state surveillance) (Solove, 2007).

The difficulties in demonstrating deterrent effects, or the absence of action that may have otherwise taken place, has driven several attempts to empirically establish their existence (Kendrick, 2013). As a result, an epistemological and methodological orthodoxy emerged asserting that if a chilling effect is to be identified, it must be independently and objectively measured (Canes-Wrone and Dorf, 2015; Penney, 2016). ³ Quantitative research focused on measuring changes in behaviour has therefore become prominent. These include the impact on Google search trends (Marthews and Tucker, 2014) or Wikipedia page views (Penney, 2016), following revelations of previously unknown government surveillance programmes. The Snowden revelations provided a useful comparator for these studies, allowing assessments of online activity before and after their public disclosure in order to examine evidence of behavioural change following new knowledge of pervasive state surveillance.

Surveys using experimental methodologies – with manipulated conditions designed to prime participants’ awareness of state surveillance before questioning them or monitoring their behaviour – have also been a popular method for attempting to demonstrate chilling effects. These studies have consistently reported that those primed with information regarding government surveillance are less likely to engage politically online (Stoycheff et al., 2018), share their views or make comments when holding what they perceive to be minority opinions (Stoycheff, 2016). Whilst these studies highlight important concerns regarding self-censorship, questions exist over the ecological validity of such insights in their translation beyond experimental conditions. Others have argued how focusing on the reduction of speech as a measurable phenomenon only provides a ‘crude proxy for chilling’ (Kendrick, 2013: 1678), obscuring the range and scope of surveillance effects.

Several qualitative studies have explored chilling effects in greater experiential detail. These have tended to emphasise experiences of disproportionately affected populations, and those – such as members of protest, opposition or political movements – whose position in society elicits a concern that they will be targeted with surveillance. Among these, studies of journalistic practices after Snowden reported changes in relationships with confidential sources (Lashmar, 2016; Waters, 2017), ‘“chilling” the flow of information’ to journalists (Lashmar, 2016: 682), and increasing vigilance and security measures adopted to protect sources (Waters, 2017).

Other research has focused on the surveillance of minority communities with the NYPD's Muslim Surveillance Program serving as a particularly stark example. Ali's (2016) ethnographic study with Muslim students and community organisers revealed notable harm and increases in self-disciplinary behaviours. These included self-censorship, particularly regarding political conversations, with a fear of informers highlighted for creating a ‘fracturing of inter-community trust’ damaging any potential for political mobilisation. Research with UK activists targeted by undercover police surveillance captured similar outcomes, including experiencing a fractured sense of reality, caused through fear and paranoia and a need to question every personal encounter and relationship (Stephens Griffin, 2020). Stephens Griffin (2020), however, prefers the term ‘derailing’ to ‘chilling’ to describe the impact surveillance had on activists, claiming that whilst surveillance created difficulties, decreased opportunities and forced shifts in focus, activists did not abstain completely. Such insights challenge binary framings of chilling effects – where subjects are either deterred or unaffected by state surveillance – to reveal a more complex continuum of adaptive responses.

One of the largest studies of social movement surveillance engaging over 70 US-based organisations revealed similar concerns regarding the breakdown of trust amongst activists as they feared infiltration (Starr et al., 2008). The study highlighted diverse effects of surveillance on activist mobilisation, as participants reported reputational damage brought by implied criminality when targeted with state surveillance. These impacts included difficulty in accessing meeting spaces and funds, the deterrence of new members and the avoidance of contentious issues. Adding further nuance to these non-individualised effects of surveillance, Starr et al. (2008) emphasised chilling of assembly as organisations struggled to survive, mobilise and maintain members once targeted. This is particularly significant given the centrality of these activities to the effective functioning of participatory democracy, and their protection under international human rights law. Freedom of expression plays a key role in facilitating the free exchange of ideas, allowing individuals to test different forms of thought and in doing so to develop their identity and their political beliefs (Joseph and Castan, 2013). As summarised by Bhagwat (2011: 978), ‘[w]ithout speech, democracy would be impossible because citizens would have no way to discuss and form their views, including their views about the conduct and competence of public officials.’ Freedom of assembly guarantees the right to gather peacefully, indoors or outdoors, and in public or private (Kudrevicius and Others v. Lithuania, 2015). As such, freedom of assembly is closely related to freedom of expression and is considered to protect that right's collective component (United Communist Party of Turkey and Others v. Turkey, 1998). Together these rights are essential to democracy, as they facilitate the emergence of new ideas, engagement with those ideas, advocacy and protest in pursuit of change and transparency and accountability in governmental decision making.

Surveillance in Zimbabwe

In Zimbabwe, state surveillance is shaped by a complex ensemble of statutes, agencies and extra procedural practices of which many are politically motivated (Munoriyarwa, 2021). Among these, the Interception of Communications Act (ICA) (2007) sets out the legal basis for state authorities to conduct communications surveillance and implicates commercial telecommunications providers in the interception and storage of private communications. Research addressing chilling effects brought by the ICA on journalism in Zimbabwe has demonstrated how metadata retention clauses included within the Act have manufactured fear among journalists, increased self-censorship and limited their access to sources. Journalists feel that the legislation is tailored to focus on activists and journalists specifically to instigate fear and paranoia and interfere with their reporting (Munoriyarwa and Chiumbu, 2020).

Reports that Chinese telecommunications company Huawei has been instrumental in enabling several States – including Zimbabwe – access private communications of political opponents and their supporters adds to these concerns (Parkinson et al., 2019). They also highlight the role of third-party actors in the provision of state surveillance capabilities and the subsequent impact on democratic participation. Similar questions of private operators are raised over tactics of ‘digital authoritarianism’ such as censorship and internet shutdowns (Mare, 2020) in addition to more targeted surveillance tools such as the Pegasus technology offered to governments by Israeli security company NSO Group (Amnesty International, 2021). In the context of shutdowns in Zimbabwe, telecoms providers have been coerced into becoming enablers rather than resistors of state desires through strategic uses of ‘lawfare’, harassment and threats of violence and imprisonment (Mare, 2020). The extent of state-owned infrastructure also further complicates the relationship between the state and private actors, as internet service providers continuing operations are subjected to the whims of that state.

Research has, however, also highlighted the important role of online platforms for organising and mobilising protest in Zimbabwe (Karekwaivanane and Mare, 2019). While not specifically focusing on the impact of surveillance on activism, Karekwaivanene and Mare (2019) suggest growing suspicions of online infiltration by government agents within online spaces. As developed in the analysis below, such fears subtly erode the democratic and deliberative potential of online spaces and deter their use for democratic participation.

Digital surveillance in Zimbabwe is also accompanied by a range of national security, or public safety focused legislation, invoked as a basis to conduct surveillance on activists and to restrict the right to freedom of assembly. These include the recently repealed Public Order and Security Act (2002) and its successor, the Maintenance of Public Order Act (2019). The latter remains extremely restrictive with any gathering over 15 people and deemed ‘political’ in nature requiring permission, which is frequently, and often selectively, denied. Such instruments have been used as a means to crack down on civil society organisers and groups such as students, with reports of arrests and violence for violating these laws (Gukurume 2019). Acts such as these also contribute to the heavy presence of surveillance on university campuses ‘deployed to whip political dissent into line’ which results in a fear of being watched and censorship permeating all aspects of university life (Gukurume 2019: 776). Such acts are characteristic of wider processes present in Zimbabwe of discipling public institutions, who are exposed to intensive surveillance in order to ensure compliance to fulfil the will of the state (Maringira, 2017; McGregor 2003). Oversight of many of these state practices is largely non-existent, and surveillance is therefore conducted with impunity. The lack of any legal framework to regulate offline surveillance makes this form of surveillance extrajudicial, in effect granting the state unfettered powers in this area. A crucial question here concerns how the unregulated and unaccountable surveillance practices arising from such statutes impact individuals to whom they are intended to target.

Methodology

Research for this paper was comprised of semi-structured interviews with key-informants holding direct experience of state-sponsored surveillance in Zimbabwe. Participants can be variously identified as human rights defenders, civil society leaders, members of opposition political parties and those working in the media. Together they represent individuals who have both been targeted by state surveillance and who also conduct work or activities which aim to hold the state and ruling party to account. They therefore occupy a unique position to provide insight into how the ruling party in Zimbabwe uses surveillance against dissenting voices and its impact on them and their activities.

These interviews were supplemented by a range of documentary sources including court documents, media reporting and documents from civil society groups which provided valuable insight into the operation and impact of state surveillance in Zimbabwe. In addition, as the data reveals, chilling effects are highly subjective and deeply personal experiences. As such, interview questions were formulated openly and broadly to access the details of participants experiences as they felt they were best described.

Accessing participants who have been subjected to state surveillance and who remain at risk of state coercion is a difficult and delicate task. Participants were approached and recruited through local networks, where access was achieved through longstanding engagement with human rights defenders in the region. In total 12 interviews were conducted, 11 match the profile described above with one additional supplementary interview with a state security agent who provided insight into the practices of state surveillance. While this is a relatively small sample, access to such communities is extremely rare and similar-sized cohorts are not unusual within this area of research. Several other potential interviewees were identified and either declined or did not respond. This is an understandable position, given how the findings presented later, and elsewhere (Gukurume, 2019), demonstrate the erosion of trust towards new individuals as a result of experiencing state surveillance. In this sense, research into surveillance chilling effects is, to an extent, chilled. Accessing participants willing to discuss their experiences becomes incredibly difficult due to suspicions created by the surveillance they have encountered and is often only rendered possible through trusted prior connections and networks. This makes the accounts of those who are willing to participate particularly valuable, as demonstrated by research using similar sample sizes that offer important qualitative insights into the resultant impacts of state surveillance (Munoriyarwa and Chiumbu, 2020; Stephens Griffin, 2020; Waters, 2017).

Ethical considerations

Participants were fully briefed regarding the aims and purpose of the research, allowing them to give their informed consent to participate. Many of the interviewees are high profile political activists in Zimbabwe who have publicly challenged the government, and been subject to physical abuse as a result. Their greater expertise on the vulnerabilities of taking part in the research was sought and incorporated into the overall ethics process with participants having further opportunity to identify any part of their interview they did not want published. Additionally, even in cases where participants noted that they were happy to speak on record – as several had in the past – the decision was made to keep accounts anonymous, enabling continuity with those who had requested anonymity. Names and identifying details have therefore been changed throughout.

Research which focuses on the targets of surveillance has the potential to increase or create further vulnerabilities for participants (Ali, 2016), as it could expose strategies of avoidance or even allow a state to better understand which tactics are most effective in controlling or silencing citizens. Careful reporting of the findings and balancing these concerns with revealing the impact of surveillance and allowing participants to have their stories heard and harm documented in a way that may enable surveillance practices to be challenged in future, was therefore of consideration throughout the entire project.

Findings

Forms of surveillance

State surveillance activities in Zimbabwe implicate complex assemblages of state and non-state actors involved in diverse and sometimes hybrid online and offline monitoring practices. As the data demonstrates, even unambiguous state surveillance activities by specialist intelligence units involve the co-option of third-party and ‘informal’ elements, sometimes amounting to a form of ‘surveillance by proxy’. Moreover, the range of potential bodies and organisations available for co-option into surveillance assemblages affects targets in complex and profound ways, indicating that chilling effects similarly comprise complex and profound forms. As such, the intricate imbrications between state and non-state, digital and non-digital forms of surveillance are key to understanding the chilling effects of surveillance in the Zimbabwean context. Before addressing these impacts, this section delineates the forms of surveillance experienced by the participants, providing the context for their accounts.

Complex collaborations: State and non-state actors

In Zimbabwe, the abduction of human rights defenders typically involves kidnappers handing victims to the police before becoming invisible to any legal scrutiny. The high profile and well-documented case of one participant is particularly illustrative of this process. After weeks of being held incommunicado, they and others finally appeared at various police stations and were charged with undergoing military training with intent to overthrow the government. They described their experience of surveillance leading up to the abduction,

The type of surveillance that I have experienced… involved State agents watching me at my house for about ten days…I was not aware of this, and the sad thing is that they ‘greased the palms’ of my helper who used to help me with the garden. So, they were communicating with him, and they spent several days coming in and out of my house, pretending at times that they’ve had a puncture just for them to be able to see what was happening at my house. I think that there had been prior arrangements made before my abduction (Participant 8)

In a more recent case, Tawanda Muchehiwa was similarly abducted and kept for more than 48 h before being handed over to the police (Sparks, 2021). Following the abduction, local investigative journalists produced evidence that a rental vehicle from a local car hire company was used to monitor and kidnap him, later a court ordered the company to turn over documentation relating to the identified vehicle (Mabuza, 2020). That no arrests for the kidnappings have been made indicates complex collaborations of formal and informal surveillance actors.

Other participants’ accounts of surveillance repeatedly pointed to the reliance of state intelligence operations on informal actors. Elaborating on this, an interviewed state security officer confirmed the existence of multiple levels of surveillance. Crucially, one reason for the widespread use of human informants is linked to labour costs in Zimbabwe: the labour of human surveillance is cheaper than in other jurisdictions and, in many cases, cheaper than engaging in more technical forms of observation. This represents an inversion of what is experienced in other jurisdictions where digital surveillance is attractive as it becomes comparatively cheaper to collect and analyse data at scale, and where human labour costs are prohibitively expensive. This brings a new dimension to the surveillance of online-hosted communications too. Rather than requisitioning data sets through bulk surveillance powers, the availability of cheap labour that can be deployed for state surveillance can infiltrate private communications in different ways, one which subverts the encryption protocols embedded in digital platforms.

Surveillance labour in the global south: The infiltration of platform communications

The infiltration of seemingly private digitally-hosted discussions challenges the distinction between surveillance of online and offline spaces. Participants raised concerns about the infiltration of civil society by unknown informants passing information to the authorities. This fear was grounded in the knowledge of increasing numbers of arrests directly resulting from contents of WhatsApp messages (Majama, 2016). This is especially common vis-à-vis ill-defined offences of insulting the president under the Criminal Code or sending malicious messages under the Postal and Telecommunications Act. In one such incident, Zimbabwe Lawyers for Human Rights (2020) reported that Abraham Baison was arrested and charged with undermining authority of the President by allegedly circulating a message on WhatsApp accusing him of ineptitude. The nature of these complaints and arrests indicates that WhatsApp and Facebook groups, while considered ‘closed’ and ‘confidential’ by their participants, are routinely infiltrated, and monitored.

Other participants emphasised how this type of surveillance in virtual spaces was made possible as people considered them private spaces shared only with friends and consequently communicated with less caution:

there is a false sense of security that comes with being online in that many people think that they are on the internet and so they will not get arrested… People think they are talking amongst friends because a WhatsApp group is usually for friends …I have an uncle who was arrested for sharing a fake prayer about the former president in a WhatsApp group a few years ago. The prayer is something that he would never have shared in a public place like a pub or church, but because he thought it was safe to share among trusted friends on a WhatsApp group, he lowered his guard and yet someone reported him to the police. (Participant 12)

Several interviewees described such infiltration of digitally-hosted forums as a reformulation and extension of pre-digital surveillance techniques, noting that, ‘It is the same method that was used in beer halls and in church gatherings…but only now they are using the same techniques in the virtual space…As soon as people get comfortable enough to speak, screenshots are taken and then used to arrest them.’ (Participant 10)

Whilst this incursion into digital spaces marks a novel form of state infiltration, it echoes Cohen's (1985) identification of the ‘net-widening’ outcomes of new control measures, constituting an extension, rather than replacement of, extant surveillance practices. For example, this type of infiltration was repeatedly cited by participants as occurring offline at gatherings and meetings, even running as deep as intimate relationships. As one activist described,

I had an intimate relationship with one of these agents, who used a false name, and he would never let me see his identity documents. At one time I saw him when there was a demonstration and he had called me before the demonstration to find out if I was going…I lied and said that I was busy at home, but I attended the demonstration, and while I was there, I saw him there. This incident made me realise that I was under surveillance and that the surveillance was at a very personal level…I started seeing shadows everywhere, even when I would go out with friends for drinks, I would leave without saying goodbye because I felt like I was constantly being watched. (Participant 11)

Such accounts reveal how this form of government surveillance asserts ancillary affects. Beyond the direct suppression of individual rights, interpersonal trust becomes a distinct casualty of such forms of surveillance, affecting the rights to freedom of expression and assembly at a group level. Additionally, as the surveillance net widens and practices extend to new contexts, a greater number of individuals become effected by such activities. Transgressions which may have previously been considered less significant to the state become easier to capture and act upon, as demonstrated by the examples of comments posted by ‘ordinary’ citizens to ‘trusted’ WhatsApp groups, widening any potential chilling effect to an increasing number of citizens.

Intentional chills

Participants’ experiences also revealed how social media information is frequently used by state actors to intimidate through direct reminders of state surveillance capabilities.

For example, during a visit to a police station to submit prior notification for a meeting of more than 15 individuals, one respondent was about her social media posts, instead of the meeting. She remarked, ‘I would be asked about my posts on social media, why I posted that opinion, what it meant and what my intention of posting them was. This is sheer evidence of online surveillance’ (Participant 4). Such practices reveal a further component of the chilling effect, one underplayed in much of the literature: how state surveillance is used to assert intentional chilling effects. This experience was common among participants and such attempts at inducing intimidation took many forms. Sometimes this was done through unsolicited anonymous messaging, such as that sent to a Harare based activist shared on social media (Figure 1). At other times, surveillance data would be produced during police interrogations, serving as a direct reminder that individuals were being watched. Additionally, such capabilities are publicly announced by the state, demonstrated by the information minister's recent remarks that they have ‘a cyber-team that is constantly on social media to monitor what people send and receive’ (quoted in Mathanda, 2021).

Figure 1.

Intentional chilling effects and WhatsApp intimidation.

Such tactics are not new, and previously publicised incidents serve as a reminder of state capabilities, reinforcing their threat. For example, in 2007, the state covertly filmed the bedroom of then Catholic Archbishop Pius Ncube, an open critic of the ZANU-PF regime. Secret cameras recorded him engaged in sexual intimacy, later broadcast on prime-time news (McGreal, 2007). Ncube resigned and stopped publicly commenting on political matters. This served as an effective warning to other clergy that had been critical of the government. Referring to the incident sometime later, the state security minister at the time stated with impunity,

be careful not to denigrate our president; we will visit your bedrooms and expose what you will be doing. We have our means of seeing things these days…So, no one can hide from us in this country. (quoted in Kahiya, 2014)

The leaking of WhatsApp group discussions to the media also serves as a public demonstration of this approach. As one interviewee reflected, this creates a sense of doubt and undermines attempts to organise,

I also think that in as much as we say that WhatsApp is a great tool for communicating it has also been a great tool for surveillance and the instances where discussions we’ve had on WhatsApp have been leaked to the media show that it is no longer a safe space to organise and mobilise

This section has highlighted the key aspects of surveillance practice in Zimbabwe. First, how surveillance is used to infiltrate ‘trusted’ spaces, extending the reach of surveillance – and thus chilling effects – beyond traditionally targeted groups and into the general public. Second, it demonstrates how publicising surveillance practices is used to intimidate and intentionally chill activist activity. The following section outlines how these activities generate new surveillance possibilities and diversify the impacts of such measures on their targets. In doing so, it reveals how these arrangements create an enduring plurality of chilling effects, demonstrating the impact on both individuals’ rights to freedom of expression and assembly, as well as the ability of groups to engage in collective democratic action.

The effects of surveillance

Participants were consistent in articulating the impact of surveillance on their lives and behaviours, both personally and professionally. Changes to behaviour evidencing a ‘chilling effect’ vary yet included: greater restraint in political conversations, increased self-censorship, limited participation on social media, amplified awareness of surroundings, eroded interpersonal trust and restricted economic participation through limited access to work and informal ‘blacklisting’. These impacts directly affect the rights to freedom of expression and assembly.

One principal driver of these chilling effects was a fear for personal safety and security:

There is a lot of fear. People can see the brutalisation that is happening to activists now and they have opted to stay quiet for their own protection. One of the reasons we have never really come to a point where we all revolt is because too few people consider it as a risk that is worth taking. Not a lot of people are willing to compromise their safety and the safety of their families, they would rather just conform and not be involved in those discussions. The risk of surveillance has cowered people into submission. (Participant 11)

This is exacerbated for those already subjected to such consequences, who reported an enduring sense of insecurity and fear; “I no longer feel safe in public areas, especially with my profile and I think other victims of surveillance feel the same way.” (Participant 6).

Another female activist who became romantically involved with someone secretly monitoring her revealed the pernicious nature of surveillance conducted in the intimate spheres of life:

As a woman activist, your whole life changes and that affects your well-being. It has been very difficult, and it has scared away most people who thought they were activists and who were coming up as activists… we are living examples of what state security can do to destroy you through surveillance. (Participant 11)

The data below further reveals the complex manifestations and, crucially, gradations, of chilling effects. While routinely characterised in binary terms – for example as an individual either being ‘chilled’ towards self-censorship or not – the data reveals subtle impacts and interplays on specific rights and behaviours that are not just individual but also vicarious and collective.

Freedom of expression

Freedom of expression is widely regarded as ‘one of the essential foundations of a democratic society and one of the basic conditions for its progress and for each individual's self-fulfilment’(Axel Springer AG v. Germany, 2012), given its centrality to human development and its essential role in securing other rights, through debate, advocacy and protest. As noted by the European Court of Human Rights:

Democracy thrives on freedom of expression. It is of the essence of democracy to allow diverse political programmes to be proposed and debated, even those that call into question the way a State is currently organised (Centro Europa 7 S.R.L and Di Stefano v. Italy, 2012)

Chilling effects on freedom of expression are therefore particularly pernicious, as they strike directly at individuals’ ability to develop their own identity, and to effectively participate in democratic life.

Common to other manifestations of surveillance chilling effects, the impacts on the right to freedom of expression are varied in form and intensity. For some, this restraint on expression involves a constant and conscious modification of speech:

I always practice what I would call self-censorship. I carefully construe how I communicate on certain issues and discuss them in a way that does not provide an opportunity for surveillance to become actionable. (Participant 1)

Restrictions on freedom of expression also extend beyond the individual and may impact how organisations participate in public debate:

On several occasions, I would find myself writing posts and deleting them several times. I started self-censoring on my individual accounts and censoring the organisation. (Participant 4)

Impacts on freedom of expression also hold downstream consequences that are often undocumented. In one case, this affected the ability to work and acquire viable income streams for those in the creative arts:

We are afraid to exercise, especially, our right to freedom of expression, which has had a great impact on our artistic work. We are afraid that if we express our opinions, we risk being labelled enemies of the State. So, we either censor ourselves and in some instances have had to turn down some clients because we are afraid that even if they are willing to pay us good money, the risk of clashing with the authorities is high. (Participant 6)

Partial mitigation strategies to counter the chill on freedom of expression have emerged. For example, some activists have taken on a ‘brokering’ role, intended to draw surveillance on them, and away from peers considered as more vulnerable:

about a third of the stuff that I tweet does not come directly from me, people send me information that they are afraid to share and ask me to open up a conversation on the issue. These include people that I interact with in civil society, government, and the private sector, who for one reason or the other feel that they cannot freely or openly speak on an issue… Surveillance has affected the way people interact with each other and the way they comment on issues. (Participant 3)

Although a courageous workaround given the circumstances, the need to adopt such measures highlights the extent of interference with the right to freedom of expression in Zimbabwe. This will have a profound impact on individual's ability to participate in democratic life and to pursue political change.

Freedom of assembly

Freedom of assembly is also regarded as central to democratic life. It ensures that individuals can organise and mobilise, come together to test or challenge ideas, gauge support for, or opposition to, different proposals, and pursue political change. As such, it is regarded as the collective exercise of the right to freedom of expression (Human Rights Committee 2020). The extension of chilling effects beyond the individual to affect the collective is a central theme throughout the data and demonstrates the wider impact of surveillance across social settings. In their most obvious form, individuals are reluctant to participate in protests, meet with particular people or groups or join specific organisations; these wider impacts directly affect the right to freedom of assembly. The data further reveals how this animates an erosion of interpersonal trust, builds a sense of vicarious victimisation, generates isolation through the loosening of community bonds and advances an overall weakening of democratic participation. As one participant stated, ‘expressing oneself on public or social fora puts you in the spotlight, and may land not only you into trouble, but also those that associate or follow you.’ (Participant 6)

Linking with the above accounts of non-state actors becoming co-opted into formal surveillance practices, the pervasiveness of surveillance exerts a toll on interpersonal trust. This, in turn, impacts personal and professional relationships and the ability to work in a normal way:

Over the past year, about three or four people that I work with have told me that they have been offered money to give information about me and my activities… I have also encountered a situation where one of my employees said that a journalist had introduced them to a state security agent who wanted to get information about me…I find I am at a point where I cannot trust my own colleagues in the sector. I cannot afford to leave my office unlocked when I go to the toilet, I cannot take my cup of coffee and leave it in the kitchen because I am always afraid that someone will poison it. (Participant 3)

The feeling that close acquaintances may be implicated in targeted surveillance practices exerted significant trauma and lifestyle alterations on individuals. Accounts of how everyday activities acquired precautionary framings were common across participants:

I think the biggest thing…is that you change your behaviour, you change where you go and who you see. Instead of mixing and mingling freely, you are always looking over your shoulders. For example, for something as stupid as going to the toilet while drinking in a pub, you must think whether you should finish off the drink or take it with me…That is the kind of thing that surveillance does. (Participant 12)

A common experience was the social ostracisation of activists, compounded by the internalisation of an outsider status. This asserts a further toll on the ability to build and maintain connections, which permeated into the closest of relationships:

I have family members who I am no longer sure about. I do not know where they stand, politically. There are also members whose actions sometimes appear to be unclear, and I find myself withdrawing from them. Even my relationship with my partner is affected, because of the perception that I am under surveillance your relationship is strained, and this also extends to my immediate family. (Participant 1)

Trust is a critical ingredient in social activism. Respondents consistently highlighted how interpersonal trust is a foundational pre-requisite for most democratic participation and advocacy work. However, the climate of surveillance combined with genuine prospects for reprisals exerts particularly far-reaching effects on the establishment and maintenance of such trust. As one participant indicated, ‘surveillance comes from all angles – from the state and even from among us’. (Participant 11)

Consequences of eroded interpersonal trust on the ability of activists to mobilise and engage in democratic debate were also consistently highlighted,

Organizing and mobilizing have been severely affected because of surveillance. Instead of organizing and mobilizing for big demonstrations, we then resorted to just organizing amongst trusted friends and having the demonstrations as a small group because we cannot trust many people getting involved and take the risk of having the demonstration stopped or exposing the main organisers. Surveillance makes you trust no one. (Participant 11)

Building on the evidence of compromised social media platform communications above, prospects of surveillance in these ‘closed’ spaces heavily depletes trust within and between activist networks, who regard each other with increasing suspicion. Many expressed how such surveillance-induced reductions in trust has curtailed their ability to participate in democratic debate and has heavily restricted the ability to organise.

Such accounts further emphasise the importance of recognising how chilling effects extend beyond individuals and assert collective impacts:

you will find threats and fearmongering in WhatsApp groups where you are reminded of arrest if you insult the President… it's been a great strategy in ensuring that the trust level is lowered in terms of the interactions online. Like sometimes you literally need to know who you are speaking to, who are your recent followers, and when did they start following you, what was happening around that time? I think just increasing the level of distrust is enough to ensure that there isn't an effective offline movement. (Participant 7)

(dis)Association/blacklisting

Those placed under surveillance are subjected to complex processes of isolation and exclusion. For example, compounding internal behavioural modifications – such as self-censorship and a reluctance to engage with others – surveillance targets found themselves blacklisted and ostracised. For example, one participant who endured a publicised abduction and torture then found their corporate and commercial contracts evaporated:

At some point, the organisation was targeted for [material] that was seen as anti-establishment… after that incident, some corporates and organizations that we used to work or partner with, now do not want to associate with us because they are afraid of being labelled enemies of the State as well. (Participant 6)

While private actors are quick to avoid those targeted (or at risk of being targeted) by the state, civil society groups also reported the need to distance themselves from such individuals or organisations. This not only erodes solidarity and collective activity among civil society groups, but underlines the wider, vicarious and collective impact of surveillance. One participant experienced a similar dynamic whereby several women's rights organisations actively distanced themselves from her organisation following her arrest:

My opinion on surveillance is that it has affected the way our audiences engage with our group. So, you will find that even the comments on our platforms have taken a significant dip after my abduction, as they risk being accused of having views that are contradictory to the government. People are now afraid to partner with us, to comment on our projects and there is a fear of adopting our doctrines. People are now afraid to associate with [us] and with me as a person. (Participant 6)

The depleted trust and disassociation between those connected to anti-state activities severely compromises the capacity of civil society groups to mobilise and to engage in advocacy work:

I think it has affected the ability to mobilise and organise to a large extent.… As a member of a social movement, you draw attention to yourself, and people begin to disassociate from you. I think surveillance has affected the effectiveness of social movements. Right now, most activists are in hiding… surveillance easily discourages active and effective participation in democratic processes in Zimbabwe. (Participant 4)

These accounts of activists and human rights defenders in Zimbabwe reveal how surveillance chilling effects extend beyond the individual to implicate civil society networks and the wider social body, undermining trust, and the ability to engage in democratic processes. At an individual level, chilling effects implicate a range of rights beyond the right to private and family life, including freedom of expression and assembly, the effects of which are also felt at a group level, implicating the effective functioning of democratic processes. These differing accounts of those targeted also demonstrate how chilling effects are not only highly diverse but are negotiated in subjective and individualised ways. The long-term effects of a surveillance-related chilling effects are potentially profound. An erosion of the rights to freedom of expression and assembly will result in the erosion of democratic life itself, and an entrenchment of the status quo.

Discussion and conclusion: Imbrications of traditional and advanced digital surveillance

State surveillance practices in Zimbabwe are diverse, expansive and permeate throughout everyday life. Drawing on the experiences of those interviewed, the findings reveal how state infiltration, and its impacts, penetrate an increasing array of physical and digital spaces, widening the pool of potential surveillance targets, and capturing an ever-greater range of acts designated subversive.

The availability of cheap labour in Zimbabwe facilitates the practice of using informants to infiltrate virtual spaces, rendering useless the technological protections of authentication and end-to-end encryption associated with the use of such platforms. Such practices not only represent a blurring of online and offline tactics but also represent an inversion of practices in other jurisdictions, where labour costs are prohibitively expensive and so digital surveillance becomes increasingly attractive as a cheaper, more convenient tool. Such practices resonate with the growing phenomenon of ‘data janitors’, employed in low wage labour economies to make sense of data rich environments, and their deployment in authoritarian contexts, such as the repression of Uighurs and Kazaks in Xingjian (Byler, 2022).

This recognition of differing practices reinforces the importance of studying wider contexts, to reveal diverse and currently unexplored manifestations and impacts of state surveillance. This broadens discussions hitherto dominated by the impact of automated large-scale monitoring programmes, such as those revealed by Edward Snowden (e.g. Penney, 2016; Stoycheff, 2016), and unpack how state surveillance is enacted, experienced and negotiated in different jurisdictions and through different subjectivities.

These findings have, however, also drawn clear parallels with research focused on the targets of state surveillance elsewhere, including the US and UK (e.g. Ali, 2016; Starr et al., 2008; Stephens Griffin, 2020). In particular, similarities in harms experienced by those who find themselves subjected to surveillance are revealed, even if negotiated in different forms, and resulting in different consequences. As in these studies, this paper demonstrates how interpersonal trust has become a distinct casualty of state surveillance practices (Ali, 2016; Stephens Griffin, 2020). Yet, such effects are exacerbated in Zimbabwe by the pervasive concern of the presence of informers in almost all contexts both online and offline. The breakdown of trust, questioning of even the most intimate relationships, the withdrawal from social life and social media, and a desire to keep circles small, exerted a discernible toll on participants as they reported increasing caution and reconsideration of every interaction and relationship amid a climate of ambient distrust. This damage to trust worked in other ways too, as peers sought distance through fear of becoming surveillance targets themselves, resulting from their association, however tenuous. This severely limits opportunities to organise, preventing democratic participation and possibilities for advocacy work. Such effects demonstrate clear and pernicious impacts of state surveillance that require mitigation through law and regulation. Moreover, such clear evidence of the profound real-life consequences for those targeted by state monitoring practices mean that surveillance chilling effects cannot be summarily dismissed as speculative, as they have in recent judicial and Congressional processes (Clapper v. Amnesty Int’l, 2013; Wikimedia Foundation v. NSA 2021). It also points towards a need to do more to challenge the ‘legal and judicial scepticism’ (Penney, 2016) of subjective experiences of state surveillance as a credible source of evidence for chilling effects that currently inform such thinking (Kendrick, 2013; Solove, 2007).

Not only does attention to this space establish the existence and impact of surveillance chilling effects but, importantly, it contributes key insights into their diverse forms and dynamics. As a result, this paper concludes that, while clear incidences of the chilling effects of state surveillance activities are identifiable, such outcomes are not deterministic, binary or necessarily predictable. Instead, chilling effects assert themselves in highly diverse ways (directly, indirectly and vicariously), become projected onto different audiences and manifest through different harms. Debates surrounding the chilling effect of state surveillance must therefore be extended beyond discussions of the presence or absence of speech, and harm must be considered more widely than through the lens of privacy or freedom of expression alone. Whilst large-scale studies of online behaviour have a place in understanding changes in behavioural trends which result from surveillance (e.g. Marthews and Tucker, 2014; Penney, 2016), such nuance, however, can only be achieved, identified and understood through recognition of the experiences of those most directly affected. Although vital, this represents a difficult task for researchers, considering the breakdowns in trust and wariness of new connections that those affected report.

In addition, as revealed through the experiences of those interviewed, and numerous documented examples, in Zimbabwe these processes and their deterrent outcomes are exacerbated, often intentionally, through media reporting of arrests flowing from online discussions, and details of how vulnerable individuals were pursued. This active and intentional cultivation of surveillance chilling effects thus reveals a further act of bad faith: the repeated denial that such effects exist at all. This critically undermines the fallacious ‘nothing to hide/nothing to fear’ justification for extended surveillance practices, often expressed by state governments and intelligence agencies the world over. That a chilling effect is intentionally induced as a deliberate strategy is itself undeniable evidence of its existence and potency and must be confronted.

The experiences of those interviewed for this paper demonstrate that surveillance chilling effects extend far beyond the suppression of individual rights. Beyond privacy, rights to freedoms of expression, and assembly, are directly engaged, reverberating such effects from the individuals targeted to impact wider society and the functioning of democratic processes. This suggests that chilling effects created by surveillance must also be viewed and interrogated as a societal issue, and not one which stops at individual modifications of behaviour. This has important consequences as we move towards more pervasive, AI-assisted, surveillance. Current debates regarding the deployment of AI technology typically focus on privacy and discrimination. This risks undermining the overall impact of surveillance. Specifically, while a chilling effect may be almost imperceptible over the short term, it may be dramatic in the long term, fundamentally undermining democratic processes. This suggests the need for further qualitative research vis-à-vis the impacts of surveillance, and an incorporation of this research into surveillance-related decision making.