Settled Habits, New Tricks: Casteist Policing Meets Big Tech in India

Ameya Bokil, Avaneendra Khare, Nikita Sonavane, Srujana Bej and Vaishali Janarthanan


May 2021

Big Tech is reinforcing and accelerating a system of caste-based discrimination in India and reinforcing the power and impunity of its police. 

It is the evening of 15 June 2019. The Station House Officer of the Kolar Road Police Station in Bhopal, the capital city of the central Indian state of Madhya Pradesh (MP), convenes his officers to announce a rise in car thefts in their jurisdiction. He states that the thieves are the local Pardhis and instructs officers to make night-time visits to the Pardhi basti (a slum colony), and pick up and detain anyone outside after dark. This goes unchallenged because the police widely believe that the Pardhis are habitual criminals, responsible for every case of house-breaking and theft. We might have found it hard to believe that the police blatantly target an entire community had we not been at the station for an entirely different reason. Every detained Pardhi would have their ‘suspicious activity’ recorded in the extensive files the police maintain on their community.

Recently, technology companies and governments are helping to digitise these police records and the surveillance of ‘suspect’ individuals who are more ‘likely’ to commit crimes. Through this rigged digital database, the Indian police force is being empowered to sustain its caste-based criminalisation of marginalised communities and continue to act arbitrarily with impunity. The digitisation of already biased police records, extensive surveillance systems, predictive policing through interlinked databases and the complete absence of a regulatory framework have led to the creation of a parallel digital caste system which denies the fundamental freedoms of specific marginalised communities.

When caste is criminality

The Constitution of India formally protects citizens’ right to equality, including the right to equal treatment before the law and freedom from discrimination on grounds of race, caste and religion. The adoption of the Constitution was a transformative moment as India became a sovereign, democratic postcolonial republic and aspired to move beyond the pervasive feudal legacy of the caste system. The entrenched nature of the caste system, applying to nearly all aspects of life, has made it hard to extirpate, making the right to equality unattainable for certain communities, as we had witnessed at the police station. The Pardhi community is one of India’s Adivasi or indigenous communities. Although formally outside strictures of the caste system, these communities are nonetheless vilified.

Several hundred communities, including the Pardhis, were branded as ‘hereditary criminals addicted to systematic commission of non-bailable offences’ under the Criminal Tribes Act (CTA) enacted by the British colonial government in 1871. Its aim was to make these communities liable to state surveillance and control in myriad ways. Since the repeal of the CTA in 1952 and the official decriminalisation of the tribes criminalised thereunder, these communities are referred to as De-notified Tribes (DNT).

The CTA was inspired by the combination of racist European criminal anthropology and the Indian caste system, which portrayed criminality as a hereditary characteristic. British colonial authorities established the police in the 1840s with the explicit objective of controlling the Indian population through force. The police failed in in its objective, so to give an appearance of order, it adopted the strategy of selective policing of certain groups. This selection hinged on a social consensus on who was a criminal, informed by the caste system.

Thus, the colonial strategy created categories of persons who were considered the ‘proper objects of policing’, principally forest-dwelling and nomadic communities (including Pardhis) whose way of life conflicted with British interests. The Pardhis are traditionally semi-nomadic hunters, whose way of life conflicted with colonial regulations on hunting, control over forests for commercial purposes and revenue collection from a sedentary population.

Consequently, surveillance and intrusive policing became a part of these communities’ daily existence: from having their names registered in permanent records, being placed in ‘reformatory settlements’ in order to undermine their nomadic cultures and facing severe restrictions on their movement as well as constantly having to report to the authorities. This system is best described by historian Radhika Singha’s explanation of the colonial policing system – ‘it was far easier to prosecute a prisoner on a charge of belonging to some ill-defined criminal collectivity than to establish individual responsibility for a specific criminal offence’.

The contemporary Indian police have continued with this legacy. First, Indian society continues to be ordered by the caste system; second, the idea of hereditary criminals still occupies the mind and structure of the Indian police, largely comprising members of oppressive castes; and third on the grounds of expediency that hold as true now as they did during British colonial times.

Thus, limited policing resources are still targeting the same communities. Even after the CTA was repealed, its legacy has endured in policing structures, practices and attitudes. Individual Indian states have adopted legal provisions concerning ‘habitual offenders’ (HOs) and maintained the surveillance systems designed under the CTA. The hereditary criminal of the past is now placed in the more palatable administrative category of the HO, which remains ill-defined and therefore gives the police vast discretionary powers. These provisions, while apparently neutral, are still selectively used against the same communities that were targeted in colonial times.

References to ‘Habitual criminals’ in the Madhya Pradesh Jail Manual, 1987, Vol 1, Part 2

A centuries-long project of predictive policing

Angrez chale gaye, police chodh gaye hamare liye’ (‘The British are gone, but they have left their police behind’), says a Pardhi woman in Bhopal, referring to the continued police discrimination her community faces. Whether in the form of indiscriminate detention, torture in custody, or economic exploitation, the everyday life of Pardhis is characterised by police violence; but because they lie at the very bottom of Indian caste society and continue to experience socioeconomic hardship, their systemic exploitation (much like their very existence) is rendered invisible. Scholars have failed to document the systemic police targeting of Pardhis and other DNT communities, while civil society’s attempts to highlight the issue have been restricted to anecdotal evidence of police brutality.

Once held at a police station, every Pardhi – children, women and men – is subjected to physical and verbal assault. ‘They know it is easy to beat a confession out of a Pardhi’, says one woman. Parents are beaten in front of their children to ‘send a message.’ Recently, two Pardhi minors were picked up from a tea-stall and temple, stripped naked and beaten. The police misled their mother about where they were, and when she tried to get them released, she too was beaten by the police and framed under false charges.

Besides the violence, police surveillance has long dispossessed Pardhis of their traditional livelihoods. Today, they depend on waste picking, begging and odd jobs like unloading rubble and clearing bushes. ‘In the entire city, there is not one person who will offer us employment in a shop or give us any salaried job’, claims one Pardhi, citing the stigma of criminality associated with DNTs. The constant police surveillance, harassment and frequent arrests hinder their ability to pursue education and steady employment.

Threatening to create a new police record or add to an existing one, the police demand large bribes (the equivalent of US$ 250–1,500) from the Pardhis. The bribe increases each time they are held in police detention or jailed. Between the bribes, bail and the lack of steady employment, a Pardhi family typically remains trapped in a cycle of perpetual indebtedness.

Police stations across India maintain registers of HOs – also called ‘history-sheeters’ – in their jurisdictions, with extensive details of their lives and daily movements. While their identification may not explicitly be based on caste, collective police action overwhelmingly identifies members of the DNT communities as HOs. These registers record their demographic details such as place of residence and caste, personal information such as age and identifying marks on the body, and ‘evidence’ of criminality: details of their habits, their method of committing crimes, their property, particulars of their associates, places they frequent, etc.

For communities such as the Pardhis, even being visibly mobile carries with it a threat of police surveillance and violence. Rana, a middle-aged Pardhi man, was stopped by traffic police for not wearing a crash helmet. When the police demanded to know his surname and caste identity, he was detained and questioned about how he had obtained a motorcycle. When his answers were deemed ‘unsatisfactory’, he was arrested. So pervasive is the fear of having one’s daily life recorded in police registers that Rana, much like other Pardhis identified as a HO, rethinks every activity of his life, including something as mundane as going to the local tea stall with friends. These records shackle the Pardhi community’s lives, freedoms and dignities.

Arguably, the most important section of the habitual offenders’ registers is an informal record that police officers must sign to attest that they have personally trailed or surveilled the HO at least once every fortnight to investigate whether s/he had (despite extensive surveillance) managed to outwit the police to commit a theft or burglary. The police in Bhopal’s Govindpura police station showed us this, which was surprising given that these registers are fiercely guarded to avoid public scrutiny.

State police regulations allow them to maintain a record of HOs in their jurisdictions, but there are no restrictions on the information they can keep in their registers and therefore no form of accountability. As a result, the police have developed their own practices, including using local informants (known as mukhbirs) to keep abreast of HOs’ daily activities and movements.

Madhya Pradesh Police Regulations detail some of the ways they surveil suspect individuals and communities

Police knowledge is ossified into algorithms

For over a century, the police have kept physical records of all cases and HOs, but these are now being digitised through the Crime and Criminal Tracking Network & Systems (CCTNS), the main and centralised system for maintaining digital records. The central government provides the core infrastructure to standardise digital data: First Information Reports (FIRs), various documents related to investigation and evidence, and the final police reports to be submitted to the courts. The CCTNS also allows geo-tagging of offences. However, for various state governments which are trying to build their own infrastructure over and above this standard, CCTNS extends far beyond digitisation into setting up a crime-mapping, analytics, and a predictive system.

A super platform and an opaque black box, the CCTNS has been designed to be the digital repository of every local police record. It is hoped that it will make policing more efficient by allowing local police stations to know about a person’s entire criminal history at the click of the mouse: cases in which the person has been accused, facial photographs, the crimes committed, the number of days held in detention, and whether the courts acquitted or convicted them.

The central government and the tech industry maintain that systems such as the CCTNS will allow for ‘objective’, ‘smart’, error-free algorithm-based detection of criminal hotspots and predictive policing. In reality since these databases are fed by the police’s centuries-long caste-based system of preventive surveillance and predictive policing (which has already determined who is a criminal and what crimes habitual criminals commit repeatedly), there is no possibility of objectivity or lack of caste bias. The CCTNS only adds a technological veneer to a caste-based policing model. While the ideological purpose matters little to local police stations, its material benefits include hours of saved time and seamless digital transmission of produced criminalities across jurisdictions.

The government and the tech industry maintain that [digital] systems will allow for ‘objective’, ‘smart’, algorithm-based detection of criminal hotspots and predictive policing. In reality since these databases are fed by the police’s centuries-long caste-based system of preventive surveillance and predictive policing there is no possibility of objectivity or lack of caste bias.

The reliance on and aspirations for predictive policing are a part of the aim for the Indian police to be among the world’s most advanced and professionalised force. The police claim that their limitations are inadequate staffing, poor technological skills and an overworked force. The goal is to have the same technological tools in every police station in India as in London and New York, to increase efficiency and technological expertise and obviate the need to recruit more police officers while also cutting workloads. Despite concerns raised about predictive policing in the UK and US with respect to racial profiling and discrimination, mass surveillance, arbitrary search and seizure, as well as the erosion of the fundamental right to privacy, the Indian police have sought to enhance predictive policing technologies such as hotspot detection and data mining.

The CCTNS is the future of India’s police registers. By 2030, it is hoped that the platform will be adequately developed to free police officers of maintaining any paper registers. Since state governments are free to tweak CCTNS as they please, several states have been collecting biometric details (iris scans, facial prints, etc.) of HOs and even first-time offenders. A senior police officer in Bhopal claimed that the CCTNS is being used in Madhya Pradesh as a repository of all criminals. CCTNS integrates various dossiers: history sheets and goonda files, fingerprints, footprints, details about family members of accused persons, etc. The details of family members are obtained for a ‘deterrent effect’, so that purported criminals do not commit further crimes. (A goonda is what the police call individuals who are more likely to commit assault or disturb public peace by indulging in general public violence and rioting. Derived from a pejorative term in Hindi that roughly translates as ‘rowdy’ or ‘hooligan’, the police surveilled such identified individuals through various Goonda Acts from as early as 1926.)

Permanent databases do not spare children either. One 16-year-old Pardhi had his details forcibly recorded (including fingerprints and photos) on a charge for which he was eventually given a suspended sentence. Creating permanent records of children, whether or not they are convicted, may be in direct contravention of the Juvenile Justice (Care and Protection of Children) Act, 2015, which espouses the principle of a ‘fresh start’ for a child. This, however, is of no concern to the local cop.

The problem here, much like in the case of traditional surveillance, is that the police operate without a clear legal framework and use ambiguity to their advantage. The Madhya Pradesh Police Regulations allow for the creation of physical databases of HOs and briefly state the manner of their surveillance. This legal framework, which has remained unchanged over decades, did not foresee today’s digital advances. Consequently, using technology for surveillance, with its associated set of new problems (permanence, security, and privacy, to name but a few), has virtually no legal basis and therefore very few constraints. In a landmark judgment in 2017, the Indian judiciary confirmed that Indian citizens have a fundamental right to privacy; however, the judiciary has yet to extend this to the question of maintenance of HO registers and databases.

The only limits to the breadth of police surveillance appear to be infrastructural constraints. To address these problems, the state of Telangana, for instance, is investing in a multi-storey centre to house its ambitious Integrated People Information Hub (IPIH), a database containing 360º profiles of every resident. Other states plan to follow suit.

The state’s inability to self-regulate its use of technology is amply demonstrated by the ham-fisted introduction of the Aadhaar, a ‘unique identification’ number linking biometric information and various databases necessary for accessing welfare programmes, setting up bank accounts, purchasing SIM cards, and paying income tax, among others until the Supreme Court directed the government to regulate and limit its mandatory use for specific public services.

No hiding from the police

A second technological advance with regard to policing is the use of closed-circuit cameras (CCTV), purportedly for national security and women’s security. A chilling incident of rape and murder that made the headlines in 2012 led to greater calls for harsh criminal laws and mass surveillance technologies to deter crimes against women. The paternalistic preoccupation with maintaining control over women’s bodies for the stated purpose of ensuring their safety has resulted in surveillance in public spaces. The Lucknow city police recently announced that they will initiate a response for women in distress based on their facial expressions observed through AI-equipped facial-recognition technology.

Most of the larger Indian cities are dotted with police CCTV (without attendant regulations) on busy streets, at intersections, and in market areas to replace in-person police surveillance. Private establishments and educational institutions in larger cities have also invested in CCTV on their premises in accordance with state regulations seeking to establish ‘public safety.’ The sinister implications of creating and maintaining such networks are obvious. Recently, during the lockdown imposed to control the COVID-19 pandemic, the Union Government allowed a hate-filled media campaign to vilify Muslims as maliciously spreading outbreaks across India.  Unsurprisingly, this culminated in the police in Madhya Pradesh Telangana using drones to surveil mainly Muslim neighbourhoods.

While CCTV-based surveillance policing has thus far largely maintained the distinction between footage from police cameras and from private cameras mandatorily installed, but accessible only after a crime has been committed, a strange third hybrid is being developed in some parts of the country. Take, for instance, Bhopal Eye, the crown jewel of the Bhopal police surveillance system. This mobile application is marketed as a citizen-policing initiative which allows the police to actively maintain a database of the number, location and range of all private CCTVs installed in the city. As part of the Bhopal Eye initiative, the local police have been ‘encouraging’ the installation of CCTV in homes and commercial establishments, even in the absence of mandatory public safety regulations.

The economic model of Bhopal Eye, in some ways, parallels mobile applications such as Uber: the financial investment of acquiring and maintaining the input units (CCTVs) for the intended output (surveillance) is made not by the organisation that built the network, but is shifted to citizens by selling to them dual myths of ever-lurking danger and the deterrent value of constant surveillance. When citizens are thus recruited, they can download the free application and log in the location coordinates of their CCTVs. The police, as the database creators, develop this network and use the data to keep track of how many of the city’s ‘private eyes’ can be harnessed for policing purposes. Apart from the lack of any regulation, little is known about the procedure used to manage Bhopal Eye, its use and its efficiency. The senior police officer credited with single-handedly constructing and initiating Bhopal Eye declined to answer our questions.

In the future, systems such as Bhopal Eye could, through both overt encouragement and tacit prejudice, facilitate the police’s surveillance reach within mixed neighbourhoods, where both affluent and working-class families reside, as well as enable heightened monitoring of ‘suspect’ individuals, such as street-vendors, in wealthy localities.

Police promotional material shared on social media sites such as Facebook to encourage Bhopal Eye registration with the tagline “No criminal will be safe if houses install cameras properly”.

Some members of the DNT communities seem amenable to constant CCTV surveillance. Rana, the man who admitted to being afraid of even going to the local tea stall due to police harassment, exclaims, ‘I wish they would actually install CCTVs everywhere so that they would know [that it is not us] who is actually committing these crimes’. The history of CCTV has taught us, however, that the purported safety-net they offer is not for all citizens. For instance, police stations are required to have cameras installed at their entrances and inside cells where the accused are detained. This was supposed to be a check on custodial violence, for which the Indian police are notorious. However, they are now careful to take the accused to a room without CCTV before assaulting them. The police influence medical examinations of the accused to obfuscate the evidence of their brutality. The police hesitate to supply CCTV video feed to those who request it, and the judiciary is not known to strictly demand it either. The reliance on cameras is thus entirely at the discretion of the local police. The police may simply not submit video evidence when it does not favour their case, while the defence might never learn that such evidence exists.

Thus, Rana’s trust in the potential benefits of CCTVs seems misplaced. A 2018 incident related to the criminalisation of cattle slaughter in the city of Khandwa illustrates the point. The police had installed CCTVs at a busy town square to check on the illegal slaughter of cattle. A Muslim man, who lived within the range of the surveillance camera, was arrested for the offence. His family repeatedly claimed that the case against him had been fabricated and that the camera’s recording would provide the proof needed to exonerate him. However, since the prosecution enjoys wide discretion in introducing evidence at trial, the family had no means of ensuring that the footage made its way to court.

What’s in it for Big Tech?

We found almost unreserved enthusiasm for technology in our interviews with the police, with officers at all levels regarding the force’s steady progress into increasingly sophisticated methods of data collection and collation as a badge of honour. The only problem, according to the officer in charge of the CCTNS at one of the local police stations, is that the tools currently at their disposal do not work well enough. Nor did we have to poke around to find the reasons for this candid embrace since we received the same reply repeatedly: technology makes policing more efficient, convenient, accessible, and accurate – basically, easier. Experience suggests that justifications of efficiency need not be set in stone. Be it the precipitous demonetisation implemented in 2016 or the gargantuan Aadhar project introduced in 2009, regimes across the political spectrum have displayed a Hydra-like ability to invent and reinvent justificatory strategies, leading one to wonder what exactly lies underneath the screeds and slogans. It is thus meaningful to ask what interests are served by the technological advances but remain obscured from public view.

One need not look far. As Usha Ramanathan, a veteran privacy activist and trenchant critic of the Aadhaar project puts unequivocally, ‘the driving force behind the [Aadhaar] project is corporate interest’. India is a staggeringly lucrative market for private corporations that produce surveillance technologies. In Bhopal, for example, security cameras alone, which are only one component of the city’s surveillance infrastructure, cost upwards of Rs. 2.5 Crores (about US $135,000), an enormous figure for a non-metropolitan city.

Indeed, a visit to the futuristic central CCTV control room at the police headquarters in Bhopal is like being at a trade fair, as a support staff member from Honeywell, the corporation behind the city-wide Integrated Video Management System project, rattles off the panoply of brands that have coalesced to create this panopticon. Honeywell landed the lucrative contract after its impressive performance at surveilling the Kumbh Mela, a Hindu pilgrimage held every 12 years, attended by millions. An eight-member team from Honeywell is constantly circulating around Bhopal’s police headquarters to provide ‘indispensable’ technical support.

The nature of Honeywell’s partnership with Bhopal police illustrates two significant ways in which private corporate interests cast an ominous shadow over the ‘public good’: not only do corporations pull the strings; they also invariably set the stage. Consider the case of Huawei, which has been selling ‘safe cities’ across the world, of which facial-recognition technology is an integral component, altering its pitch to appeal to diverse potential patrons.

Similarly, Honeywell’s offering sees public safety as a function of constant, blanket surveillance, whereby everybody is suspect unless observed otherwise – a vision that it then volunteers to execute and helps the law-enforcement agency navigate. Honeywell’s role is currently limited to providing the police with CCTV footage when a crime is committed and they request assistance to discover who the criminal is or track where they might have fled. However, for the police and Honeywell, this is only its secondary, instrumental purpose. The inherent purpose of Honeywell’s CCTV footage is that it compels public self-discipline and self-surveillance, and therefore reduces crime.

Another case in point is the proliferation of facial-recognition-based AI technologies that are being marketed to police departments across the country as the ‘must have’ crime-fighting tools. According to reports, the Surat City Police has a picture intelligence unit that relies on Nippon Electrical Company’s proprietary NeoFace technology and vehicle number-plate recognition to track persons of interest. During one of our interviews with a police officer, he alluded to the city’s population as ‘clients.’ This language reveals what is not being openly stated, which is that the state is motivated more by being the customers of shiny curios than acting as the custodian of the citizens’ civil rights.

The construction of a digital caste system

The dream of Digital India was sold to the Indian population by the Narendra Modi government soon after being elected in 2014. A key component of this dream is to build an e-governance model. The need to use technology has been furthered through the myth that tech is neutral in the prevention of crime and curbs the problem of human bias, when all that these systems do is essentially digitise the casteist targeting of communities through the nebulous category of HOs. There is scant reflection on what such a digitised caste system implies, who is responsible for designing it and how it reproduces and reifies hierarchies that are inimical to the criminal justice system.

The goal of efficiency makes no attempt to dislodge the traditional principles of policing: who is kept under surveillance and why remain constant across time and space. Technological advances merely sharpen the blade of police discretion and further entrench its operational biases. This is already a departure from the supposed relationship between technology and law enforcement in countries where police body cams are at least expected to correct implicit bias.

Indeed, in the US, there have long been arguments about the attempt to shroud racist systems under the cloak of objectivity. The historian Khalil Muhammad demonstrated how a ‘racial data revolution’ in the nineteenth century marshaled science and statistics to make a ‘disinterested’ case for white superiority. The results of the 1890 census showed that African Americans were disproportionately imprisoned, but rather than interpreting this as a symptom of systemic inequities, the data was understood to be ‘objective, colour blind, and incontrovertible’. In this way, crime statistics became the foundation upon which the myth of Black inferiority was constructed.

Likewise, the datasets and models used in newer, tech-based systems are not objective representations of reality. The employment of new technologies that reflect and reproduce existing inequities – but that are promoted and perceived as more objective or progressive than the discriminatory systems of a previous era – often hides, speeds up, and even deepens discrimination, while appearing to be neutral or benevolent.

A spate of new applications allows the police to access and update information from their mobile phones. The more one asks them about the number of databases the police ‘maintain’ and their regulatory frameworks, the more obvious it is that even the police have been unable to keep up with the mushrooming of private technologies in their operations. It is unknown how many online databases there are, how information is uploaded onto them, the nature and breadth of this information – and what is the ultimate aim of assembling this large archipelago of digital database infrastructures for police surveillance. These multiple applications, software and databases may (for now) exist in silos, even if they regularly cross-pollinate information. However, they are all so close to each other that they can all be easily interlinked to build dossiers of personal information on all citizens and allow more pervasive institutional profiling, which can then be used to justify differential treatment in schooling, employment, housing, etc., particularly for those belonging to marginalised communities who are identified as HOs.

Thus, when the Pardhi community says that the police have the entire history of every member of their community, this is no exaggeration. It is clear that the real aim of surveillance and the unchecked powers it gives to the police is to maintain political hegemony and a very strict, hierarchical social order. Thus, surveillance policing allows for the marriage of profit-making corporations and authoritarian regimes. The social control they seek to maintain is, in turn, in accordance with the casteist social control already enforced by police surveillance.

A Case for Police Accountability

We need to go beyond resisting the introduction and use of surveillance technologies to question, if not overhaul, the very ethos of policing in India, because it has already perpetuated dangerous predictive policing on the bodies of the marginalised even without these technologies. Unfortunately, the Indian state is increasing its excessive reliance on the institutions of policing to respond to various crises. Recently published data, for instance, documents the state’s heavy reliance on policing and colonial-era sedition law to quell dissent on issues ranging from Kudankulam Nuclear Power Plant to the Citizenship Amendment Act, 2019. Each new cause opens up new avenues for criminalising and policing marginalised communities.

New justifications for increased police surveillance will keep multiplying as the state seeks to gain a tighter hold over the social order. Each instance of a perceived threat to the ‘internal security’ of this social order is fertile ground for intensifying surveillance mechanisms, as has been observed most recently in governments’ response to citizens’ protests. The habitual offender is to the city what the ‘anti-national’ dissenter is to the country – an inveterate, antisocial element from whom society needs protection.

Activists, lawyers, students, and a cultivated category of urban Naxals are now all beginning to be at the receiving end of this uninhibited state surveillance and records. In February 2021, The Washington Post reported that civil rights activist Rona Wilson’s laptop had been hacked into for surveilling and planting false documents implicating him as an enemy of the state. While the hacker’s identity is unknown, it is reported that the hacker (an individual or organisation) had extensive resources at their disposal. In 2019, it was reported that the Israeli firm NSO Group’s spyware tool Pegasus was used to surveil journalists and human-rights defenders. The NSO Group admitted that it sells Pegasus exclusively to governments and law-enforcement agencies.

Surveillance-based policing to address violence against women is the latest addition to the police’s range of responsibilities. Although feminist and women’s movements, among others, have questioned the failure of criminal law to address violence against women, the questions of caste-based criminality, policing and intersectionality have largely remained absent from this conversation. The caste-based construction of criminality makes women from marginalised communities the most vulnerable. They suffer the greatest violence but receive no support from the criminal justice system, because its very structures allow the police to perpetrate such violence in the first place.

A parade of women walking from their dorms to various factories sidestepping sewage and construction. Credit: Andrea Bruce / NOOR

Feminist and civil rights movements in India have essentially failed to question the very ethos of casteist policing in the country. Pardhi women recount instances of police harassment when they are at work segregating waste. If the police find anything valuable in their waste-collection bag, they are accused of having stolen it and are dragged to the police station. Some women said that their own jewellery and other items are confiscated and recorded as evidence of theft. In two instances where Pardhi women had committed suicide because of police harassment, the state relied on such ‘criminal antecedents’ to portray these women as criminals and grant the police impunity – as if being a criminal justified being a victim of police violence. The instances of violence against Pardhi women are rendered invisible through these constructed narratives of criminality, granting the police complete discretion and impunity.

We therefore need to focus our resistance on the very cause of the problem of disproportionate police targeting of marginalised Adivasi communities either through technological surveillance or other means: police discretion and impunity. As gatekeepers of the criminal justice system, the police determine who become the subjects of this system. Investing in building police accountability is the first step in tackling the culture of discretion and impunity that has become synonymous with law and order in India.

By underscoring and drawing attention to forms of coded inequality, not only must we challenge the social dimensions of technology, but also work against the construction of a parallel digital caste system that essentially intensifies the traditional caste system. At present, those whose bodies are subjected to violence by the carceral system are forced to seek justice from the same system. In the US, the Black Lives Matter (BLM) movement has highlighted the dangers of carceral imagination, the racist systems of policing and the need to invest in non-retributive forms of justice.

A major BLM demand is to defund the police who are designed to criminalise Black communities. Lessons from the movement can be adopted to challenge caste-based oppression in India through policing. This would begin by building a discourse of the casteist nature of policing through advocacy, research and community organising, alongside an active investment in cultivating an imagination of a transformative justice, one that is not designed to prosper on the bodies of marginalised communities, whatever their caste – the eradication of which is another, much larger, struggle.


Ameya Bokil, Nikita Sonavane and Srujana Bej are young researchers and lawyers working at a nascent criminal justice collective in Bhopal, India called the Criminal Justice and Police Accountability Project (CPAProject). Ameya Bokil is a human rights researcher examining overcriminalisation and police powers, Nikita Sonavane has worked and written extensively on issues of sexual violence, indigenous self governance and anti-discrimination law, Avaneendra Khare is an advocate currently based in Bhopal, Vaishali Janarthanan is a final year law student and works in organisational capacity building, and Srujana Bej has pursued field based research on the impact of Aadhaar-based biometric authentication on access to socio-economic rights. @srujana_bej @CPAProjectIndia

Filename: BRA2016003C-0053-bokil-1/Credit: © Stanley Greene / NOOR

Caption: India, Andhra Pradesh, 2012, Para military routing the Protest demonstration. The conflicts between paramilitary forces and Maoists is centered on the Red corridor and goes from the south-east of India to the north-east and gather 7 regions of India – Karnataka, Andhra pradesh, Chattisgarth, orissa, jharkhand, Bihar and West Bengal. IG Handels: @stanleygreene Twitter Handels: @noorimages

Filename: BRA2016003C-0053-bokil-2 / Credit: ©Andrea Bruce / NOOR

Caption: India, Bangalore, 09 August 2016, A parade of women walk from their dorms to various factories in Bangalore every morning, sidestepping sewage and construction. India is trying to transform itself into a manufacturing economy, and for this, it must find workers. In this case, workers have been brought here at government expense from remote villages, from a population that has never before been thrust into the economy: unmarried women. IG Handels: @andreabruce  Twitter Handels: @abruce_noor


If you are enjoying State of Power, please consider making a donation. No amount is too small. Your contribution will ensure our independence and sustain the Transnational Institute into the future.