Authored By: Gunjan Ukey
Kalinga University
I. INTRODUCTION
On 1 January 2023, about 1.4 million persons who depended on this system for their food were denied their monthly ration and could not meet their most basic needs for survival — not for a single day, but for an entire month. There was no personal investigation, no written record of the decision, no public discussion, and no committee to adjudicate the issue. Those individuals had no warning. One day their Aadhaar records were active; the next, following the workings of an automated system that nobody could fully explain, they were declared duplicates or inactive. This was not a one-time, random mistake — it had been going on for at least two months. Over the past decade, algorithm-based systems have come to play an increasingly central role in Indian governance. This includes benefits ranging from scholarships to subsidies delivered to about 912 million people through Aadhaar linkage, in a program known as Direct Benefit Transfer, or DBT.
II. CONSTITUTIONAL FRAMEWORK
The next subject in this discussion is Article 14, the constitutional guarantee of equality before the law. Article 14 of the Constitution states that “the State shall not deny to any person equality before the law or the equal protection of the laws within the territory of India.”
Note: sources cited in this section — Internet Freedom Foundation, Algorithmic Governance in India: A Constitutional Audit (IFF 2024); Unique Identification Authority of India, Annual Report 2022–23 (UIDAI 2023); Central Board of Direct Taxes, Faceless Assessment Scheme: Guidelines and Procedure (Ministry of Finance 2020); Software Freedom Law Centre India, Facial Recognition Technology and Civil Liberties in India (SFLC 2023).
This guarantee has since been developed through subsequent judgments of the Indian judiciary, including the landmark case of Maneka Gandhi v. Union of India (1978), in which Justice Bhagwati held that equality could not be merely a formal rule — it had to provide a reasonable framework for the implementation of policy. He further held that arbitrary action must be the exception, since equality and arbitrariness are mutually exclusive: any action that violates the former cannot be consistent with Article 14 unless it adheres to a rational classification scheme that pursues a legitimate object. Arbitrariness itself was held to be an affront to equality, which the Court has recognised as part of the Basic Structure of the Indian Constitution.
III. STATE DEPLOYMENTS OF ALGORITHMIC GOVERNANCE IN INDIA
A. Aadhaar-Linked Welfare Distribution
The Aadhaar Act provides the statutory framework for the issuance of an Aadhaar number and its use in delivering subsidies, benefits, and services — including, among other things, any benefit, subsidy, or service for which eligibility is based on a residency requirement. The procedures for capturing and collecting demographic and biometric information are set out in the 2016 Act, and have been found by the Supreme Court to be substantially the same as those under the earlier 2011 framework.
B. Faceless Assessment Scheme Under the Income Tax Act, 1961
This subsection concerns the Faceless Assessment Scheme, under which no human officer is directly involved in allocating or carrying out tax assessments. The scheme has been administered by the Central Board of Direct Taxes since September 2020.
Note: sources cited in this section — E.P. Royappa v. State of Tamil Nadu (1974) 4 SCC 3, para 85 (Bhagwati J); Maneka Gandhi v. Union of India (1978) 1 SCC 248, para 56 (Bhagwati J); Ajay Hasia v. Khalid Mujib Sehravardi (1981) 1 SCC 722, para 16; Shayara Bano v. Union of India (2017) 9 SCC 1, para 101 (Nariman J); Olga Tellis v. Bombay Municipal Corporation (1985) 3 SCC 545, para 32; Paschim Banga Khet Mazdoor Samity v. State of West Bengal (1996) 4 SCC 37, para 9.
C. Facial Recognition Technology in Law Enforcement
Facial recognition technology is being deployed unevenly across jurisdictions and used for different purposes. In India, the Delhi Police is using it for surveillance, the Telangana Police has incorporated it into a new surveillance initiative, and the Uttar Pradesh Police is using it for traffic enforcement. [Flagged for verification: the original draft claimed that “the U.S. has not adopted Facial Recognition technology in major law enforcement” and referenced a U.S. system called “AMBIS.” Neither claim could be verified — U.S. federal law enforcement, through the FBI’s Next Generation Identification–Interstate Photo System (NGI-IPS), has used facial recognition since 2011. Please confirm the intended comparative point and the correct system name before publication.]
D. Predictive Policing Systems
Few Indian states are without some plan to employ predictive policing systems — many are already using them, or are in the process of implementing them. The full infrastructure required is not yet in place, but such systems can be built on the foundation already laid by the Crime and Criminal Tracking Network and Systems (CCTNS).
IV. CONSTITUTIONAL VIOLATIONS ARISING FROM AUTOMATED DECISION-MAKING
A. Violation of the Requirement of Reasoned Orders
Under Indian administrative law, an authority is obliged to give reasons for its decisions — a duty grounded in the principles of natural justice and in Article 14 of the Constitution. The Supreme Court addressed this requirement in S.N. Mukherjee v. Union of India. [Flagged for verification: the direct quotation attributed to this case in the original draft could not be independently confirmed and has been removed pending source-checking; the underlying legal point about reasoned orders is well established independent of this specific quotation.]
Note: sources cited in this section — State of Punjab v. Ram Lubhaya Bagga (1998) 4 SCC 117, para 12; Shreya Singhal v. Union of India (2015) 5 SCC 1, para 93 (Nariman J); Aadhaar (Targeted Delivery of Financial and Other Subsidies, Benefits and Services) Act 2016, s. 7; Justice K.S. Puttaswamy (Retd.) v. Union of India (2017) 10 SCC 1 (Nine-Judge Constitution Bench).
B. Violation of Audi Alteram Partem
These cases make clear that the principle of audi alteram partem applies strictly to administrative decisions taken by any authority when determining a person’s rights or entitlements. This principle has since become embedded in Indian administrative and constitutional law jurisprudence.
C. Violation of the Right Against Arbitrary Classification
Later judgments held that a classification is not ultra vires Article 14 provided it satisfies a recognised constitutional test. In Ram Krishna Dalmia v. Justice S.R. Tendolkar (1958), the Supreme Court set out that test: any classification must rest on an intelligible differentia that distinguishes the persons or things grouped together, and that differentia must bear a rational nexus to the object sought to be achieved by the law in question.
V. NATURAL JUSTICE AND ALGORITHMIC DECISION-MAKING
A. The Doctrine of Natural Justice in Indian Law
It is well established that the right to have one’s case decided in accordance with the principles of natural justice forms part of the implicit content of the Constitution. Beyond its express provisions, the Constitution embodies unwritten principles that have come to be universally recognised as “natural justice,” and the Supreme Court has repeatedly held that these principles are reflected throughout our constitutional framework. Justice Bhagwati’s reasoning in Maneka Gandhi v. Union of India remains the leading illustration of this point.
Note: sources cited in this section — Income Tax Act 1961, s. 144B (as amended by the Finance Act 2021); Central Board of Direct Taxes Notification No. 60/2020, dated 13 August 2020; Parliamentary Standing Committee on Communications and Information Technology, Report on Facial Recognition Technology (Lok Sabha Secretariat 2023); Vrinda Bhandari and Renuka Sane, ‘Towards a Privacy Framework for India in the Age of the Internet’ (2018) 10 NUJS Law Review 1, 18.
B. The Right to an Explanation
It would be inconsistent, then, not to read a “right to know the reasons” into the procedures that process citizens’ data. If a citizen has a right to know the reasons behind a judicial decision, a system that reaches decisions through automated processing ought equally to provide an explanation that makes its reasoning intelligible.
C. The Human-in-the-Loop Requirement
There must, therefore, always be a human decision-maker somewhere within the system — someone able to exercise discretion or override the algorithmic process — in any case touching on a person’s fundamental rights, where the algorithmic decision affects that person’s welfare, personal liberty, or right to livelihood. This requirement can be grounded either in the guarantee of fair procedure under Article 21, or in the protection against arbitrary action under Article 14.
VI. DISCUSSION
A. The Accountability Vacuum
An algorithm can fail in any number of ways without there being an obvious means of assigning blame, requiring the flaw to be corrected, providing a path to seek damages, or establishing a mechanism for review and appeal.
B. The Bias Problem and Article 14
Algorithms are generally trained on historical data, and that is the only data they can process. The difficulty is that much of this historical data reflects existing patterns of racial, social, and economic inequality. It should be no surprise, then, that an algorithm trained on the state’s past decisions tends to reproduce those same patterns rather than correct them.
Note: sources cited in this section — S.N. Mukherjee v. Union of India (1990) 4 SCC 594, para 35; Union of India v. Mohan Lal Capoor (1973) 2 SCC 836, para 26; A.K. Kraipak v. Union of India AIR 1970 SC 150, para 15.
This same reasoning was engaged by the Supreme Court in Shreya Singhal v. Union of India (2015), in the context of Section 69A of the Information Technology Act and Article 19; the concerns the Court raised there about unreviewable, opaque state action apply with equal force to algorithmic governance more broadly.
C. The Digital Personal Data Protection Act, 2023 (DPDP Act) and Its Limitations
The DPDP Act 2023 is the principal legislation governing digital personal data protection in India. It is organised into nine chapters comprising forty-four sections, with particular emphasis on three categories of data: significant data, personal data, and sensitive personal data. The Act provides individuals with three core rights: the right to access information about their data, the right to correct or delete that data, and the right to file a complaint.
VII. CONCLUSION
This essay should be read against the backdrop of an Indian state moving steadily toward greater algorithmic governance. That context must also engage with a broader question: how ill-suited our existing legal system is to address these new forms of governance technology. This assessment reveals a real inadequacy in how current constitutional principles are understood and applied to automated decision-making. It also shows why the solution many scholars propose — a citizen’s right to an explanation — is, as this essay has argued, insufficient to address the underlying structural and ethical problems. This essay has therefore argued for the creation of an independent body, the Independent Algorithm Audit Authority (IAAA), to monitor, review, and, where necessary, sanction the state’s use of algorithms. In a world increasingly governed by algorithmic processes — shaping welfare distribution, tax assessments, law enforcement, and social control, among much else — the assumption that algorithms are unbiased, objective predictors of outcome is, as this essay has shown, a dangerously simplistic fantasy.
Note: sources cited in this section — Ram Krishna Dalmia v. Justice S.R. Tendolkar AIR 1958 SC 538, para 11; Union of India v. Tulsiram Patel (1985) 3 SCC 398, para 90; Mohinder Singh Gill v. Chief Election Commissioner (1978) 1 SCC 405, para 8; Digital Personal Data Protection Act 2023 (Act No. 22 of 2023); Satender Kumar Antil v. Central Bureau of Investigation (2022) 10 SCC 51, para 71.
REFERENCE(S):
A. Table of Cases
A.K. Kraipak v. Union of India AIR 1970 SC 150
Ajay Hasia v. Khalid Mujib Sehravardi (1981) 1 SCC 722
E.P. Royappa v. State of Tamil Nadu (1974) 4 SCC 3
Justice K.S. Puttaswamy (Retd.) v. Union of India (2017) 10 SCC 1
Maneka Gandhi v. Union of India (1978) 1 SCC 248
Mohinder Singh Gill v. Chief Election Commissioner (1978) 1 SCC 405
Olga Tellis v. Bombay Municipal Corporation (1985) 3 SCC 545
Paschim Banga Khet Mazdoor Samity v. State of West Bengal (1996) 4 SCC 37
Ram Krishna Dalmia v. Justice S.R. Tendolkar AIR 1958 SC 538
Satender Kumar Antil v. Central Bureau of Investigation (2022) 10 SCC 51
Shayara Bano v. Union of India (2017) 9 SCC 1
Shreya Singhal v. Union of India (2015) 5 SCC 1
S.N. Mukherjee v. Union of India (1990) 4 SCC 594
State of Punjab v. Ram Lubhaya Bagga (1998) 4 SCC 117
Union of India v. Mohan Lal Capoor (1973) 2 SCC 836
Union of India v. Tulsiram Patel (1985) 3 SCC 398
B. Table of Legislation
Indian Legislation:
Aadhaar (Targeted Delivery of Financial and Other Subsidies, Benefits and Services) Act 2016 (Act No. 18 of 2016)
Bharatiya Nagarik Suraksha Sanhita 2023 (Act No. 46 of 2023)
Constitution of India 1950, arts 14, 15, 19, 21
Digital Personal Data Protection Act 2023 (Act No. 22 of 2023)
Income Tax Act 1961 (Act No. 43 of 1961)
Information Technology Act 2000 (Act No. 21 of 2000), ss 69A, 79
Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules 2021
International Legislation:
Regulation (EU) 2016/679 (General Data Protection Regulation) [2016] OJ L119/1
Regulation (EU) 2022/2065 (Digital Services Act) [2022] OJ L277/1
Regulation (EU) 2024/1689 (Artificial Intelligence Act) [2024] OJ L1689/1
Executive Order 14110 on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence (30 October 2023) 88 Federal Register 75191
C. Books
Basu DD, Commentary on the Constitution of India (9th edn, LexisNexis 2014)
Eubanks V, Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor (St Martin’s Press 2018)
Jain MP, Indian Constitutional Law (8th edn, LexisNexis 2018)
Kamath N, Law Relating to Computers, Internet and E-Commerce (6th edn, Universal Law Publishing 2020)
Pasquale F, The Black Box Society: The Secret Algorithms That Control Money and Information (Harvard University Press 2015)
Sathe SP, Administrative Law (7th edn, LexisNexis 2004)
D. Journal Articles
Arun C, ‘AI and the Administrative State’ (2022) 14 NUJS Law Review 1
Bhandari V and Sane R, ‘Towards a Privacy Framework for India in the Age of the Internet’ (2018) 10 NUJS Law Review 1
Ramanathan U, ‘Aadhaar and the Architecture of Exclusion’ (2018) 53(14) Economic and Political Weekly 12
E. Reports and Official Documents
Central Board of Direct Taxes, Faceless Assessment Scheme: Guidelines and Procedure (Ministry of Finance 2020)
India Justice Report 2025 (Tata Trusts 2025)
Internet Freedom Foundation, Algorithmic Governance in India: A Constitutional Audit (IFF 2024)
Parliamentary Standing Committee on Communications and Information Technology, Report on Facial Recognition Technology (Lok Sabha Secretariat 2023)
Software Freedom Law Centre India, Facial Recognition Technology and Civil Liberties in India (SFLC 2023)
UN Special Rapporteur on the Right to Privacy, Report on Privacy and Artificial Intelligence (UN Doc A/HRC/48/31, 2021)
Unique Identification Authority of India, Annual Report 2022–23 (UIDAI 2023)
F. Online Sources
Internet Freedom Foundation, accessed 16 April 2026
Ministry of Electronics and Information Technology, accessed 16 April 2026
Software Freedom Law Centre India, accessed 16 April 2026
Unique Identification Authority of India, accessed 16 April 2026
