INTRODUCTION
The proliferation of artificial intelligence (AI) and algorithmic decision-making systems across governance, the judiciary, and public administration has ushered in an era in which automated systems determine bail, welfare eligibility, tax scrutiny, and social credit. The constitutional implications of such opaque systems, often dubbed “black boxes”remain deeply underexplored in Indian jurisprudence. This blog contends that the right to an explanation for algorithmic decisions is an incipient constitutional right nestled within Article 21 of the Constitution of India, and that its non-recognition constitutes an emerging and urgent lacuna in Indian law.
The Supreme Court’s landmark judgment in Justice K.S. Puttaswamy (Retd.) v Union of India recognised informational privacy as a fundamental right under Article 21.[1] Yet, the right to privacy alone is insufficient when automated systems, operating without human deliberation, make life-altering decisions about citizens. The question that now presses upon constitutional courts is: does a citizen possess a fundamental right to know why an algorithm rendered a particular decision about them?
THE CONSTITUTIONAL FRAMEWORK
Article 21 of the Constitution guarantees that no person shall be deprived of their life or personal liberty except according to procedure established by law.[2] The Supreme Court has, over decades, read into this provision an ever-expanding catalogue of rights, from the right to livelihood to the right to health. The philosophical foundation for a right to explanation already exists within the Puttaswamy framework, which identified “decisional autonomy” as a core dimension of the right to privacy.[3] Autonomy, meaningfully exercised, presupposes that a citizen can understand, contest, and seek redress against decisions that affect them. An unexplained algorithmic verdict fundamentally subverts this autonomy.
Furthermore, the due process doctrine, imported into Article 21 through Maneka Gandhi v Union of India, mandates that any procedure that deprives a person of liberty must be fair, just, and reasonable.[4] An opaque algorithmic process, by definition, cannot satisfy the test of fairness when its internal logic- its weights, training data, and inference pathways is withheld from the very person it adjudicates upon. This opacity constitutes a structural denial of natural justice.
THE STATUTORY POSITION AND ITS INADEQUACY
The Digital Personal Data Protection Act, 2023 (DPDPA) represents India’s first comprehensive data protection legislation.[5] While it recognises data principals’ rights to access information and seek grievance redressal, it conspicuously omits any right to explanation for automated decisions. The earlier Personal Data Protection Bill, 2019 had proposed modest restrictions on automated processing,[6] but even those tepid safeguards were ultimately excised from the enacted legislation.
This stands in sharp contrast with the European Union’s General Data Protection Regulation (GDPR), which under Article 22[7] confers upon data subjects the right not to be subject to solely automated decisions that produce significant effects, and mandates that meaningful information about the logic involved must be furnished.[8] The EU framework thus operationalises algorithmic accountability at the statutory level, a standard India has yet to match.
THE RIGHT TO EXPLANATION AS A FUNDAMENTAL RIGHT
The argument for constitutionalising the right to explanation proceeds on three interlocking grounds.
First, proportionality: the Puttaswamy judgment adopted a four-fold proportionality test for any curtailment of privacy legitimacy, necessity, proportionality stricto sensu, and procedural safeguards. Algorithmic decisions that deprive citizens of liberty or entitlements, without offering any explanatory mechanism, cannot satisfy the procedural safeguards limb of this test. The absence of explanation is therefore prima facie unconstitutional.
Second, freedom of speech and expression: in Shreya Singhal v Union of India, the Supreme Court struck down Section 66A of the Information Technology Act for its vagueness and chilling effect on speech.[9] Algorithmic systems that take adverse action against citizens, removing content, flagging accounts, or blocking services without disclosing their rationale, produce a structurally identical chilling effect. A right to explanation is thus also anchored in Article 19(1)(a).
Third, equality and non-discrimination: machine learning algorithms trained on historically biased data risk perpetuating and amplifying structural discrimination against marginalised communities. Without a right to explanation, these communities cannot identify, challenge, or litigate discriminatory algorithmic outputs. The denial of explanation therefore also implicates Articles 14 and 15[10] of the Constitution.
THE WAY FORWARD
India’s NITI Aayog has itself acknowledged the imperative of responsible and explainable AI.[11] However, policy aspirations must be translated into justiciable rights. The following measures are urgently required.
The DPDPA must be amended to expressly recognise a right to meaningful explanation for automated decisions that produce significant legal or similarly significant effects. This includes not merely the fact of automation, but the principal factors and logic that drove the decision. Academic scholars have long argued that the “black box society” represents a fundamental threat to democratic accountability, a threat that courts and legislatures must jointly address.[12]
Additionally, sector-specific regulations in banking, criminal justice, and taxation must mandate algorithmic impact assessments and require deploying entities to maintain explainability logs that are accessible to affected individuals and, where necessary, to adjudicating courts.
CONCLUSION
The intersection of artificial intelligence and constitutional law is no longer a speculative frontier; it is a live and pressing constitutional question. As algorithmic systems increasingly exercise quasi-governmental power over citizens’ lives, the constitutional framework must evolve to ensure that the promise of Article 21 is not hollowed out by technological opacity. The right to an explanation for algorithmic decisions is not a technocratic luxury; it is a constitutional necessity, and Indian courts would do well to recognise it as such before the lacuna deepens beyond repair. Such recognition would strengthen constitutional accountability.
Author(s) Name: Aditi Kulkarni (Brainware University)
References:
[1]Justice K.S. Puttaswamy (Retd.) v Union of India (2017) 10 SCC 1.
[2] The Constitution of India, art 21.
[3]Puttaswamy (n 1) [310] (Chandrachud J).
[4]Maneka Gandhi v Union of India (1978) 1 SCC 248.
[5]Digital Personal Data Protection Act 2023, s 2(n).
[6]Personal Data Protection Bill 2019, cl 3(28).
[7] The Constitution of India, art 22.
[8]General Data Protection Regulation (EU) 2016/679, art 22.
[9]Shreya Singhal v Union of India (2015) 5 SCC 1.
[10] The Constitution of India, art 14 and 15.
[11] NITI Aayog, Responsible AI for All: Operationalising India’s National Strategy (NITI Aayog 2021) <https://www.niti.gov.in/sites/default/files/2021-02/Responsible-AI-22022021.pdf> accessed 27 June 2026.
[12]Frank Pasquale, The Black Box Society (Harvard University Press 2015) 8.