INTRODUCTION
One of the most sophisticated and invasive forms of biometric monitoring in modern digital cultures is facial recognition technology (FRT). Unlike traditional identification systems that rely on paperwork or physical verification, FRT extracts unique biometric markers from facial features and converts them into mathematical representations that enable automated identification, verification, and behavioral tracking across various environments.1The move from targeted monitoring to ubiquitous, ambient, and invisible observation woven into daily life is a paradigm shift in the nature of governmental authority, as researchers point out.²The ability of FRT to identify individuals at a distance, without physical contact or consent, positions it at the centre of global debates on privacy, autonomy, and legality.2 FRT is at the center of international discussions on privacy, autonomy, and legality because of its capacity to identify people remotely without physical touch or consent.
Over the past ten years, there has been an unparalleled growth in the scope and intent of FRT deployments in India. Retrospective criminal investigation databases and real-time crowd scanning are only two examples of how state agencies, such as the Delhi Police, Hyderabad Police, and several State Crime Records Bureaus, have incorporated FRT into their policing procedures.3 Concurrently, as part of “seamless travel” and identity-verification systems under Digital India programs, public infrastructure locations like airports, train stations, toll booths, and metro stations have started incorporating FRT.4 A major step toward developing India’s first centralized biometric surveillance architecture is the National Crime Records Bureau’s proposal for a nationwide Automated Facial Recognition System (AFRS), which aims to link various databases, including passport records, driver’s license records, and criminal records.
The fast spread of these technologies occurred against a backdrop of legislative inadequacies, notably due to the lack of a comprehensive data protection framework until the Digital Personal Data Protection Act, 2023 (DPDPA). Prior to the enactment of the DPDPA, monitoring activities worked mostly through executive notices, police tenders, and Smart City Mission instructions lacking formal underpinning. Scholars have underlined that such unchecked expansions of biometric infrastructures risk transforming welfare-oriented and administrative databases into instruments of social control through “function creep”—the progressive repurposing of data for aims outside its initial scope.5 Even with the establishment of the DPDPA, wide exemptions for state authorities create worries regarding unfettered discretionary rights in biometric processing.6
Because FRT has the potential to change the relationship between the state and its citizens, it raises constitutional concerns. At the center of this discourse lies the landmark ruling in K.S. Puttaswamy v. Union of India (2017), where a nine-judge bench of the Supreme Court unequivocally affirmed privacy as a fundamental right under Article 21. According to the Court, privacy is fundamental to human dignity, autonomy, freedom of choice, bodily integrity, and informational self-determination.7 By making it possible to identify someone without their knowledge or agreement, to track their physical movements, and to aggregate sensitive personal data from several state databases, FRT directly violates these values. As legal scholars have argued, the ability to identify individuals ubiquitously lowers the conceptual boundary between public and private areas, thereby weakening the constitutional promise that an individual retains pockets of anonymity even in public contexts.8
Further problems arise with algorithmic bias, misidentification, and unfair enforcement. Multiple empirical investigations reveal that commercially deployed FRT systems display disproportionately high error rates for women, children, and those with darker skin tones.9 The convergence of FRT with pre-existing structural disparities increases the potential of selective state targeting and discriminatory policing in the Indian context, which is marked by caste-based and religious profiling. Articles 14 and 15, which forbid discrimination and arbitrariness, present serious constitutional issues. Constitutional law from E.P. Royappa, Maneka Gandhi, and Navtej Johar has evolved to understand equality as a guarantee against arbitrary, opaque, or disproportionate state action—standards that FRT often fails to achieve due to its opacity and algorithmic unpredictability.10
Similarly, the chilling effect of frequent biometric surveillance weakens liberties under Article 19(1)(a), (b), and (d). Sociological study demonstrates that individuals adjust their conduct when subjected to surveillance, becoming less eager to join in protests, political rallies, or rebellious actions.11 The deployment of FRT at public assemblies, religious congregations, and political events thereby creates an environment where fundamental freedoms are exercised under forcible monitoring.
The article addresses these difficulties in the larger theoretical framework of constitutionalism, state authority, and surveillance. FRT’s architecture—characterised by centralized databases, real-time analytics, and cross-linking of identity systems—equips the state with unequaled capacities of social monitoring. Legal academics warn that without tight regulation, FRT can function as an instrument of “digital authoritarianism,” facilitating the surveillance of political dissidents, journalists, protestors, and minority groups.12 The Supreme Court’s insistence in Puttaswamy on the “necessity and proportionality” criterion provides a robust framework to analyze the constitutionality of such technology.13
FACIAL RECOGNITION TECHNOLOGY: NATURE, FUNCTIONING & CONSTITUTIONAL RISKS
Facial Recognition Technology (FRT) must be analysed not just as a technological tool but as a structural component of emerging state surveillance infrastructures. Its constitutional concerns derive from the way technology transforms individuals’ biological traits into data points capable of ongoing tracking, comparison, and analysis. FRT systems function through a multilayered pipeline: gathering of facial images, extraction of nodal points (such as inter-ocular distance, jawline curvature, skin texture patterns), transforming the extracted data into feature vectors, and subsequent matching with templates stored in centralised databases.14 By transforming identity into a continually traceable biometric item, this conversion of the human face into a digital identification modifies the relationship between the individual and the state.
What sets FRT apart from traditional CCTV-enabled monitoring is the automation it incorporates. Scholars remark that surveillance evolves from observational watching to computational recognition, offering unprecedented scalability and granularity in population monitoring.15 For this reason, legal theorists refer to FRT as a “power-multiplying” technology since its real-time identification capability greatly increases the state’s ability to monitor public behavior, control dissent, and gather population-level intelligence.[3] These technical capabilities pose constitutional concerns about arbitrariness, opacity, and excessive delegation of authority in situations like India, where numerous state authorities operate under broad discretionary police powers.16
Verification and identification are the two main modes of operation for FRT systems. Verification (1:1 comparison), which is frequently used for digital authentication or access control, compares a live facial image to a recorded biometric data. Identification (1:N comparison), however, matches an unknown face against big databases to determine identification. The latter kind is constitutionally more intrusive since it permits the state to perform massive population scanning without any reasonable suspicion or warrant requirement.
Technical studies emphasize the probabilistic character of FRT algorithms. The quality, diversity, and representativeness of training data have a major impact on identification accuracy since systems rely on machine-learning models trained on big datasets.17Systems produce disproportionate false positives for specific demographic groups, particularly women, children, and darker-skinned populations, when training datasets are unbalanced, as they frequently are. These concerns have been carefully documented in global literature.18In the Indian setting, where caste, religion, and regional stratification dominate policing patterns, the application of probabilistic methods without controls risks exacerbating structural inequities.
Constitutional Risks Embedded in FRT Architecture
Mass Surveillance and the Erosion of Anonymity
FRT enables the potential of real-time mass surveillance, where individuals may be identified and followed across cities utilizing integrated CCTV networks. Unlike traditional surveillance, FRT does not only observe action; it establishes identification and trajectory. As Lyon notes, this marks the transition from “sousveillance” to “dataveillance”—a system in which individuals generate constant informational traces.19 In India, the incorporation of FRT into Smart Cities Mission infrastructure causes concerns about the establishment of biometric surveillance corridors able to perform real-time population tracking.
The Supreme Court in Puttaswamy declared that privacy extends even in public areas, guaranteeing the individual’s right to be let alone and freedom from unwelcome surveillance.20Therefore, the “reasonable expectation of privacy” doctrine is violated by FRT’s elimination of anonymity, particularly when it is implemented without legal permission.
Absence of Consent and Autonomy Violations
A fundamental component of informational privacy is consent. FRT deployment—especially by police—rarely includes actual consent. Individuals cannot opt out of being scanned in public locations, nor are they alerted that biometric processing is occurring. In Puttaswamy, the Court emphasized voluntary control over personal data and acknowledged informational liberty as fundamental to Article 21.21 Non-consensual extraction of biometrics, thus, breaches decisional autonomy and bodily integrity.
FRT systems educated on non-representative datasets promote algorithmic discrimination. Buolamwini and Gebru’s seminal study, earlier cited, revealed error disparities across gender and race categories, pointing out error rates for dark-skinned women reaching up to 34.7 per cent, significantly higher than for lighter-skinned men.22 Supra worries become especially serious in India, where sociological studies document that state enforcement disproportionately affects Dalit, Muslim, Adivasi, and lower-income communities.23 Implementing FRT within such structural inequalities risks strengthening discriminatory policing, violating Articles 14 and 15.
Mission Drift and Function Creep
The concept of “function creep” refers to the progressive extension of a technology beyond its intended purpose. Ramanathan’s examination of Aadhaar shows how interdepartmental connections allowed systems designed for welfare distribution to spread into financial monitoring and law enforcement.24The same phenomenon is likely with FRT. Large state warehouses of biometric data may be formed without legal authorization by discreetly integrating databases designed for “public safety,” crowd control, or missing persons with criminal intelligence systems.
Data Security, Irreversibility, and Vulnerability
Biometric breaches are irreversible: fingerprints, iris patterns, and facial templates cannot be modified once compromised. Several Indian police databases lack strong encryption, audit trails, or deletion procedures, according to reports from the Internet Freedom Foundation.25 The risk of identity theft, profiling, and unauthorized mass surveillance is increased when large biometric repositories are paired with weak governance mechanisms.
The Structural Consequences for Constitutional Democracy
When seen collectively—mass surveillance capabilities, non-consensual biometric extraction, algorithmic opacity, discriminatory potential, and absence of legislative safeguards—FRT poses structural risks to India’s constitutional order. The foundation of constitutional democracies is the idea that state authority must be reasonable, transparent, and limited. By providing the state with a widespread informational edge over citizens, surveillance technologies reverse this relationship. Zuboff defines such systems as enabling “instrumentarian26 power,” where behavioural data becomes a tool of predictive government.
FACIAL RECOGNITION & THE FUNDAMENTAL RIGHT TO PRIVACY UNDER ARTICLE 21
The constitutional framework of privacy, which the Supreme Court acknowledges as essential to liberty, autonomy, and dignity, must be the primary framework used to analyze facial recognition technology (FRT). In Justice K.S. Puttaswamy (Retd.) v. Union of India, the Court ruled that Article 21 protects the geographical, decisional, and informational aspects of privacy.27 All of these aspects are impacted by FRT since it permits automated identification and continuous surveillance without knowledge or consent. Scholars say that the very threat of extensive biometric surveillance dramatically changes the relationship between the state and citizen, establishing structural power imbalances.28
Without a thorough legal framework, FRT has been rapidly implemented in India, especially by police departments, airports, and Smart City infrastructures. The Internet Freedom Foundation’s Panoptic Report demonstrates that most FRT deployments are based on executive tenders rather than regulations established by Parliament.29 According to Puttaswamy, this violates the first limb of the four-part privacy test (legality, legitimate objective, proportionality, and procedural safeguards).30The Court emphasized that, because sensitive biometric data is immutable, it must be strictly necessary and narrowly tailored.31 FRT, by contrast, harvests biometrics passively from everyone within camera range—violating autonomy and breaching constitutional boundaries.
Reasonable Expectation of Privacy in Public Spaces
The assertion that there is “no privacy in public spaces” is constitutionally erroneous. In Puttaswamy, the Supreme Court declared that privacy follows the individual, not the place.32 In a similar vein, Daniel Solove contends that even in cases when data gathering takes place in public, privacy concerns result from data aggregation and analysis.33FRT collapses anonymity by transforming real-time mobility into identifiable data traces, so compromising spatial and informational privacy simultaneously.
The “panopticon effect,” which is extensively addressed in surveillance theory, postulates that people who are constantly under observation self-censor and alter their behavior.34 Although the loss of privacy under Article 21 is the basis for this, Article 19 concerns are also raised.
Non-Consensual Biometric Extraction and Autonomy Violations
Non-consensual biometric acquisition directly breaches decisional autonomy, a basic component of privacy. In Suchita Srivastava v. Chandigarh Administration, the Court declared that bodily integrity and decisional autonomy are protected under Article 21.35FRT systems in India—whether utilized by the Delhi Police, Hyderabad Police, DigiYatra, or Smart City platforms—collect face data without notice, choice, or consent.
Woodrow Hartzog adds that “invisible design choices that obscure data collection undermine the moral legitimacy of consent.”36 Because people cannot choose not to have their facial data collected, FRT is the most severe form of passive surveillance.
Informational Privacy and the Irreversibility of Biometric Harm
Biometric information is unchangeable, especially face templates. The NCRB’s proposed Automated Facial Recognition System (AFRS) aims to integrate disparate databases like passports, driving licenses, criminal records, and missing persons registries.37The Supreme Court acknowledged in Puttaswamy (Aadhaar) that biometric misuse causes irreversible harm, necessitating heightened scrutiny.38This aggregation violates principles of necessity and data minimization.
Comparative jurisprudence also supports this position. In Bridges v. South Wales Police, the UK Court of Appeal concluded that real-time FRT without statutory safeguards violates privacy rights.39 The GDPR and the EU AI Act identify facial biometrics as “special category data,” subject to rigorous processing constraints. India has no corresponding statutory guardrails.
PROPORTIONALITY ANALYSIS & DUE PROCESS
The Supreme Court’s four-part proportionality test, which states that any restriction on fundamental rights must satisfy suitability, necessity, least restrictive means, and balancing, must be used to evaluate the constitutional validity of facial recognition technology (FRT). This test was established in Modern Dental College v. State of Madhya Pradesh.40 FRT, as deployed in India, fails each limb of this test. First, appropriateness needs a rational relationship between the measure used and the purpose sought. Nevertheless, there is no factual proof that FRT improves law enforcement results; instead, research conducted worldwide repeatedly reveal high false-positive rates, especially for minority groups, women, and people with darker skin tones, making FRT an unreliable policing tool.41 Second, non-biometric alternatives—like targeted surveillance, human verification, or warrant-based tracking—can accomplish the same goal without subjecting entire populations to biometric monitoring, which violates the requirement of necessity, which states that no less intrusive alternative be available. Third, FRT considerably above the constitutional barrier of minimally harming rights because it amounts to indiscriminate, dragnet surveillance of people regardless of suspicion, violating the least restrictive means requirement. Furthermore, in the balancing stage, FRT imposes significantly severe burdens on privacy, autonomy, and democratic freedoms, while the State’s justifications—typically presented as efficiency in crowd management, traffic control, or administrative convenience—are relatively weak and inadequate to outweigh the rights infringements involved.
The constitutional standards of due process, justice, and procedural safeguards found in Article 21 are closely related to the proportionality analysis. Since Maneka Gandhi v. Union of India, fairness, non-arbitrariness, and rationality have become requirements for every official action that affects individual liberty.42 FRT profoundly contradicts these standards. Firstly, there is no legislative oversight, unlike telephone interception which is regulated by the procedural safeguards established in PUCL v. Union of India; FRT, in contrast, operates entirely through executive directions, police tenders, or Smart City Mission protocols, none of which comply with the constitutional requirement of “law” for privacy intrusions. Second, police departments frequently refuse to reveal algorithmic accuracy, false-positive rates, training datasets, retention periods, or vendor-specific design features. This prevents citizens from examining the technology or contesting its implementation.43Third, FRT raises significant evidentiary issues because courts may mistakenly treat FRT-generated matches as neutral or scientifically objective despite substantial research demonstrating algorithmic bias and error, undermining fair trial rights under Articles 20 and 21. Fourth, persons subjected to FRT are given no notice or opportunity for remedy, meaning they neither know when their biometrics are taken nor have any way to contest incorrect matches, unlawful surveillance, or database inclusion.
When considered collectively, these shortcomings—opaqueness, unreliable outputs, lack of remedies, and lack of a constitutional framework—show that the current FRT deployments in India violate both the procedural and substantive aspects of Article 21. The system consequently fails constitutional scrutiny not just because it is disproportionate but also because it works without the safeguards essential to ensure fairness, accountability, and democratic legitimacy.
FACIAL RECOGNITION & EQUALITY: ARTICLES 14 AND 15
Facial Recognition Technology (FRT) raises serious constitutional concerns under Articles 14 and 15, both of which ban arbitrary and discriminatory state action. Equal protection under the law and equality before the law is guaranteed by Article 14. In E.P. Royappa v. official of Tamil Nadu, the Supreme Court famously stated that “arbitrariness and equality are sworn enemies”, finding that any official action distinguished by arbitrariness is necessarily unlawful.44 This view was reinforced in Maneka Gandhi v. Union of India, when the Court stated that Article 14 bans unreasonableness in whatever manner.45 FRT is a technology that can result in systemic discrimination, arbitrary policing outcomes, and disproportionate costs on vulnerable communities, all of which violate constitutional equality requirements when considered in the context of this doctrine.
Algorithmic Bias and Discriminatory Error Rates
FRT systems routinely result in higher error rates for women, people with darker skin tones, members of ethnic minorities, and children, according to empirical research.[3] In their fundamental work, Buolamwini and Gebru reported error discrepancies as high as 34.7% for darker-skinned women, compared with less than 1% for lighter-skinned men.46 These discrepancies are not only technical; they translate directly into constitutional harm when exploited by law enforcement.
Algorithmic bias exacerbates historical injustices in India, where Muslim, Dalit, and Adivasi populations are disproportionately targeted by policing methods. As Bhasin’s research indicates, minority groups already experience heightened surveillance and profiling in India’s criminal justice system.47 Therefore, employing FRT in policing situations risks formalising technological casteism and integrating existing prejudices into automated systems – a clear violation of Article 14.
Manifest Arbitrariness in FRT Deployments
In Shayara Bano v. Union of India, the Supreme Court established the notion of manifest arbitrariness, ruling that any governmental action or law that is arbitrary, illogical, or devoid of a sufficient governing principle is unconstitutional.48
FRT deployments in India implemented through executive orders, policing tenders, and Smart City protocols lack:
• statutory backing,
• accuracy standards,
• retention limits,
• independent audits, and
• judicial oversight
Such unbridled discretion satisfies the criteria for blatant arbitrariness. When FRT is selectively utilized during political protests or religious meetings, it exacerbates discriminatory impacts and violates the equal protection guarantee under Article 14.
Discriminatory Profiling and Predictive Policing
When connected with predictive policing technologies, FRT can create automatic risk scores, suspicion labels, and behavioural profiles. Ferguson points out that in the US, minority neighborhoods have been disproportionately targeted by predictive police techniques, which has strengthened systemic racism.49
India’s integration of FRT with the Crime and Criminal Tracking Network System (CCTNS) threatens similar discriminatory profiling. If FRT-enabled policing disproportionately impacts specific communities based on caste, religion, or area – it breaches Article 15(1), which outlaws’ discrimination on these grounds.50
Exclusion from Welfare and Essential Services
FRT is increasingly used in welfare schemes for identity verification, attendance, and delivery of public benefits. Algorithmic inaccuracies especially for the elderly, persons with disabilities, women, and children can result in denial of benefits.
In K.S. Puttaswamy (Aadhaar), the Court held that welfare systems must avoid exclusion errors, especially when dealing with vulnerable populations.51 FRT’s inherent instability in biometric recognition creates a high risk of exclusion, violating Article 14’s guarantee of equal access to state resources.
Failure of the Reasonable Classification Test
Two requirements must be met for a classification to be constitutionally legitimate under Article 14’s classical doctrine:
a distinct, understandable difference, and
a logical connection between the desired goal and the differentia.52
Both are absent from FRT-based classifications like “suspect,” “match,” and “non-match”: a comprehensible basis (due to excessive false positives), and
reasonable nexus (because results don’t consistently support justifiable police objectives).
As a result, FRT violates both the reasonable categorization doctrine and the arbitrariness requirement.
STATE POWER, SURVEILLANCE & CONSTITUTIONAL LIMITS
Facial Recognition Technology (FRT), which enables the transition from targeted, suspicion-based police to ubiquitous, real-time biometric surveillance, is one of the most important extensions of state authority in contemporary India. In Justice K.S. Puttaswamy (Privacy), the Supreme Court stressed that unchecked surveillance can reverse the basic relationship between citizens and the state, transforming people from subjects with rights into objects of constant observation..53 By enabling the State to identify, track, and catalog people in public areas without a warrant, suspicion, or legal authorization, FRT exacerbates this worry. Integrated with the Crime and Criminal Tracking Network System (CCTNS), Smart City infrastructure, and large-scale CCTV grids, FRT builds a parallel surveillance architecture capable of tracking citizens at unprecedented size and granularity. Along with the advent of new technology, there is a constitutional risk associated with the structural restructuring of state authority when executive convenience replaces constitutional accountability.
This disparity is made worse by the lack of specific regulation governing FRT. Unlike telephone interception, which is regulated by the procedural safeguards provided forth in PUCL v. Union of India, FRT functions completely through policing tenders, municipal contracts, and executive announcements.54 These instruments do not satisfy the legality criteria set in Puttaswamy, which stipulates that limits on privacy must be backed by clear statutory authority. Intrusive technologies that restrict rights cannot be authorized by executive orders. Furthermore, the dearth of clear retention periods, transparency mandates, judicial warrants, algorithmic audits, or parliamentary oversight allows FRT to expand through “function creep” — an occurrence recorded by scholars where systems generated for narrow administrative purposes over time evolve into tools of mass surveillance.55 This unrestrained growth violates the doctrine of limited government embedded in Articles 14, 19, and 21.
FRT also bears major democratic implications. Studies on surveillance consistently show that people change their behavior when they are in danger of being watched all the time, reducing their participation in protests, political gatherings, religious organizations, and dissident neighborhoods. This chilling effect is obviously a threat to the freedoms of speech, assembly, and movement protected by Article 19. The normalization of real-time surveillance is an existential threat to democratic life in a constitutional democracy that depends on free speech, active political engagement, and thoughtful discourse. FRT threatens fundamental freedom by rendering public engagement visible, recordable, and potentially punished, especially in circumstances where policing is affected by political and social biases. Any FRT deployment must adhere to the constitutional standards of legality, necessity, proportionality, and procedural safeguards in light of these concerns. Nevertheless, the majority of current systems don’t meet these standards. The State has not proven that population-wide real-time biometric identification is necessary or that less invasive options, such judicial warrants or focused surveillance, are insufficient. The State must use the least restrictive method possible in accordance with the proportionality doctrine established in Modern Dental College v. State of Madhya Pradesh.56 FRT, however, is the most intrusive conceivable form of identification, not the least restrictive. The constitutional breach is thus structural, not incidental.
In light of these deficiencies, India requires a comprehensive FRT regulation law that sets strict boundaries on state surveillance. Such legislation must: (i) prohibit real-time or bulk FRT surveillance except under judicial warrant; (ii) mandate algorithmic transparency, accuracy audits, and public reporting; (iii) impose strict data-retention limits; (iv) create an independent oversight authority with powers to investigate misuse; and (v) ensure enforceable remedies for wrongful identification or unlawful surveillance. The law must further incorporate procedural safeguards similar to those governing interception, as mandated in PUCL, to ensure that fundamental rights are not subordinated to executive discretion. Without these protections, FRT will keep growing in an unaccountable and opaque manner. In conclusion, both the technological design of FRT and its ability to change the structure of governmental authority present constitutional problems. By deteriorating anonymity, limiting autonomy, promoting biased policing, and chilling democratic engagement, FRT risks turning India from a constitutional democracy regulated by the rule of law into a surveillance-oriented state governed by executive convenience. A rights-based regulatory framework is therefore crucial. Only via statutory monitoring, judicial control, transparency measures, and severe constraints on data processing can India ensure that technology promotes constitutional principles rather than subverting them.
*****
Footnotes
1. Woodrow Hartzog & Evan Selinger, “Facial Recognition is the Perfect Tool for Oppression,” The MIT Press Reader (2021). See also Joy Buolamwini & Timnit Gebru, “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification,” Proceedings of Machine Learning Research (2018).
2. Shoshana Zuboff, The Age of Surveillance Capitalism (PublicAffairs, 2019) p. 331; David Lyon, “Surveillance, Power and Everyday Life,” Oxford Research Encyclopedia of Politics (2018); Clare Garvie, Alvaro Bedoya & Jonathan Frankle, The Perpetual Line-Up: Unregulated Police Face Recognition in America, Georgetown Law Center on Privacy & Technology (2016)
3. Vidushi Marda and Shivangi Narayan, ‘Data Systems in Policing: Examining the Use of Facial Recognition in India’ (2020) 55(34) Economic & Political Weekly 23.
4. National Crime Records Bureau, Request for Proposal: Automated Facial Recognition System (AFRS) (Ministry of Home Affairs, 2019).
5. Usha Ramanathan, ‘Aadhaar: From Welfare to Surveillance’ (2014) 49(13) Economic & Political Weekly 43.
6. Gautam Bhatia, ‘The Digital Personal Data Protection Act, 2023: Constitutional Concerns’ (2023) https://indconlawphil.wordpress.com accessed 3 December 2025.
7. Justice K.S. Puttaswamy (Retd.) v. Union of India (2017) 10 SCC 1
8. Daniel J Solove, ‘A Taxonomy of Privacy’ (2006) 154(3) University of Pennsylvania Law Review 477.
9. Joy Buolamwini and Timnit Gebru, ‘Gender Shades: Intersectional Accuracy Disparities in Commercial Facial Recognition’ (2018) 81 Proceedings of Machine Learning Research 1.
10. E.P. Royappa v. State of Tamil Nadu (1974) 4 SCC 3; Maneka Gandhi v. Union of India (1978) 1 SCC 248; Navtej Singh Johar v. Union of India (2018) 10 SCC 1.
11. Shoshana Zuboff, The Age of Surveillance Capitalism (Profile Books 2019) 321.
12. Samuel Warren and Louis Brandeis, ‘The Right to Privacy’ (1890) 4 Harvard Law Review 193; Also see—Adam Gill, ‘Digital Authoritarianism and Biometric Control’ (2021) 14(2) Journal of Human Rights & Technology 67.
13. Internet Freedom Foundation, Panoptic: A Study of Facial Recognition Surveillance in India (2022).
14. Woodrow Hartzog, Privacy’s Blueprint: The Battle to Control the Design of New Technologies (Harvard University Press 2018) 87.
15. David Lyon, Surveillance Studies: An Overview (Polity Press 2007) 41.
16. Shoshana Zuboff, The Age of Surveillance Capitalism (Profile Books 2019) 311.
17. Jenna Burrell, ‘How the Machine “Thinks”: Understanding Opacity in Machine Learning Algorithms’ (2016) 3 Big Data & Society 1.
18. Joy Buolamwini and Timnit Gebru, ‘Gender Shades: Intersectional Accuracy Disparities in Commercial Facial Recognition’ (2018) 81 Proceedings of Machine Learning Research 1.
19. Lyon (n 2) 87.
20. Supra note 6
21. Ibid.
22. Buolamwini and Gebru (n 5).
23. Anuradha Bhasin, ‘Predictive Policing and Minority Surveillance in India’ (2021) 14(2) Journal of Law and Society 77.
24. Usha Ramanathan, ‘Aadhaar: From Welfare to Surveillance’ (2014) 49(13) Economic & Political Weekly 43.
25. Internet Freedom Foundation, Panoptic: A Study of Facial Recognition Surveillance in India (2022).
26. Zuboff (n 3) 321.
27. Justice K.S. Puttaswamy (Retd.) v. Union of India (2017) 10 SCC 1.
28. Shoshana Zuboff, The Age of Surveillance Capitalism (Profile Books 2019) 311.
29. Internet Freedom Foundation, Panoptic: A Study of Facial Recognition in India (2022).
30. Supra note 6
31. K.S. Puttaswamy (Aadhaar) v. Union of India (2018) 1 SCC 809.
32. Ibid.
33. Daniel J Solove, ‘A Taxonomy of Privacy’ (2006) 154(3) University of Pennsylvania Law Review 477.
34. David Lyon, Surveillance Studies: An Overview (Polity Press 2007) 87.
35. Suchita Srivastava v. Chandigarh Administration (2009) 9 SCC 1.
36. Woodrow Hartzog, Privacy’s Blueprint (Harvard University Press 2018) 112.
37. Supra note 30
38. National Crime Records Bureau, AFRS Tender Document (2019).
39. R (Bridges) v. Chief Constable of South Wales Police [2020] EWCA Civ 1058.
40. Modern Dental College and Research Centre v. State of Madhya Pradesh (2016) 7 SCC 353.
41. Joy Buolamwini and Timnit Gebru, ‘Gender Shades: Intersectional Accuracy Disparities in Commercial Facial Recognition’ (2018) 81 Proceedings of Machine Learning Research 1.
42. Maneka Gandhi v. Union of India (1978) 1 SCC 248.
43. Internet Freedom Foundation, Panoptic: A Study of Facial Recognition Surveillance in India (2022).
44. E.P. Royappa v. State of Tamil Nadu, (1974) 4 SCC 3.
45. Maneka Gandhi v. Union of India, (1978) 1 SCC 248.
46. Joy Buolamwini and Timnit Gebru, ‘Gender Shades: Intersectional Accuracy Disparities in Commercial Facial Recognition’ (2018) 81 Proceedings of Machine Learning Research 1.
47. Anuradha Bhasin, ‘Predictive Policing and Minority Surveillance in India’ (2021) 14(2) Journal of Law and Society 77.
48. Shayara Bano v. Union of India, (2017) 9 SCC 1.
49. Andrew Guthrie Ferguson, The Rise of Big Data Policing (New York University Press 2017) 89.
50. State of Gujarat v. Mirzapur Moti Kureshi Kassab Jamat, (2005) 8 SCC 534
51. Supra note 6
52. State of West Bengal v. Anwar Ali Sarkar, AIR 1952 SC 75.
53. Supra note 6
54. People’s Union for Civil Liberties (PUCL) v. Union of India, (1997) 1 SCC 301.
55. Usha Ramanathan, ‘Aadhaar: From Welfare to Surveillance’ (2014) 49(13) Economic and Political Weekly 43.
56. Modern Dental College and Research Centre v. State of Madhya Pradesh, (2016) 7 SCC 353.