Artificial Intelligence And Legal Accountability In India - A Socio-Legal Perspective On Emerging Regulatory Challenges
Author: LawSphereX
Category: Technology, Cyber & Privacy
Published:
Updated:
ARTIFICIAL INTELLIGENCE AND LEGAL
ACCOUNTABILITY IN INDIA: A SOCIO-LEGAL
PERSPECTIVE ON EMERGING REGULATORY
CHALLENGES
Joylin Jacob Y1, Ancy Dorshia M2 & Aalan Joe Edwin3
ABSTRACT
Artificial Intelligence (AI) is reshaping governance, commerce, public services and social life in India. Its deployment across healthcare, finance, criminal justice, social welfare distribution, surveillance, and digital platforms raises pressing questions of accountability, fairness, transparency, and human rights. This paper examines the socio-legal landscape of AI accountability in India: it surveys existing statutes and policy, examines judicial responses relevant to automated decision-making and data protection, identifies contemporary legal and socio-ethical challenges, and proposes a forward-looking regulatory architecture adapted to Indian constitutional and developmental realities.
India already possesses building blocks relevant to AI governance the Information Technology Act, sectoral laws, and (since 2023) the Digital Personal Data Protection Act plus policy articulations such as NITI Aayog’s National Strategy on AI. Landmark Supreme Court jurisprudence on privacy (Justice K. S. Puttaswamy) and intermediary liability (Shreya Singhal) shape the constitutional contours of AI regulation. However, gaps remain: (1) no single statutory regime explicitly governs algorithmic accountability or high-risk AI systems; (2) enforcement architecture for auditing, transparency, and contestability of automated decisions is weak; (3) marginalized groups face disproportionate risks from biased models and opaque platforms; and (4) state uses of AI for surveillance and welfare delivery present due-process and discrimination risks.
This paper argues for a layered regulatory model combining (i) a core federal statute addressing automated decision-making, explainability, and redress; (ii) strong data protection and sectoral AI standards; (iii) independent 1 JOYLIN JACOB Y, Research scholar in Bharath Institute of Higher Education and Research at school of law/ Assistant Professor in Mugil college of Law, Tamil Nadu 2 ANCY DORSHIA, Research scholar in Dr MGR Educational and Research Institute at school of law/ Assistant Professor in Mugil college of Law, Tamil Nadu
3 AALAN JOE EDWIN, Assistant Professor in Mugil college of Law, Tamil Nadu
auditing and a certified regulator with rule-making and enforcement powers; and (iv) judicial and administrative remedies that preserve fundamental rights. The recommendations emphasize proportionality, human-in-the-loop safeguards, public-interest algorithmic impact assessments, capacity building, and multi-stakeholder governance. The paper concludes that legal accountability for AI in India must blend constitutional rights protection with regulatory agility to enable socially beneficial innovation while preventing systemic harms.
Keywords: Artificial Intelligence (AI); Legal Accountability; Socio-Legal Framework; Algorithmic Governance; Data Protection; Privacy; Transparency; Bias and Discrimination; Digital India; Constitutional Rights; Fundamental Rights; Puttaswamy Judgment; Shreya Singhal Case; Automated Decision-Making; Regulatory Challenges; AI Ethics; Governance and Technology; Human Rights; Digital Transformation; India.
INTRODUCTION
Artificial Intelligence (AI) has emerged as one of the most transformative technologies of the twenty-first century, reshaping economies, governance structures, and human interaction on a global scale. In India, the integration of AI into critical sectors such as healthcare, education, finance, agriculture, and governance reflects both the promise of technological progress and the peril of inadequate regulatory preparedness. The socio-legal implications of AI are profound, implicating fundamental constitutional values, human rights, privacy, transparency, and accountability. As algorithmic decision-making increasingly mediates access to welfare benefits, credit, employment opportunities, and justice delivery, questions arise as to who bears responsibility when automated systems fail, discriminate, or harm individuals.
India’s rapid digitalisation, driven by initiatives like Digital India, Smart Cities Mission, and the proliferation of the Aadhaar digital identity system, has created fertile ground for the diffusion of AI technologies. Governmental agencies now employ machine-learning models for predictive policing, tax enforcement, and welfare targeting, while private corporations deploy AI for automated credit scoring, recruitment, and content moderation. However, the accelerated pace of adoption has outstripped the evolution of a coherent legal and ethical framework. Unlike the European Union’s proposed AI Act or the OECD’s established AI principles, India currently lacks a unified legislative regime that addresses algorithmic accountability, bias mitigation, and explainability.
From a socio-legal standpoint, the integration of AI into the Indian legal system engages multiple dimensions — constitutional law, administrative law, data protection, and human rights jurisprudence. Landmark decisions such as Justice K. S. Puttaswamy v. Union of India (2017) 4established privacy as a fundamental right, while Shreya Singhal v. Union of India (2015) 5reaffirmed freedom of expression in the digital domain. These judicial pronouncements form the bedrock upon which an accountability framework for AI must be built. Yet, the absence of explicit statutory recognition of algorithmic harms leaves victims of automated injustice with limited redressal mechanisms. The emergence of AI also presents unique challenges to India’s pluralistic society: algorithmic bias risks entrenching existing social hierarchies, while opaque decision systems threaten procedural fairness.
The need of the hour is a balanced approach one that harnesses AI’s developmental potential while embedding legal safeguards to ensure human dignity, transparency, and fairness. This paper therefore explores how India can construct a socio-legal architecture of accountability that aligns technological innovation with constitutional morality and democratic oversight. Artificial Intelligence (AI) systems broadly, computational methods that perform tasks normally requiring human intelligence are being integrated into Indian public administration and private services with speed. AI promises efficiency, predictive capability, and scale, but brings risks of systemic bias, opaque decision-making, automated exclusion, and novel privacy intrusions.
Socio-legal accountability requires law and institutions that allocate responsibility for harms, enable remedies, and constrain misuse while permitting beneficial deployments. This paper situates AI accountability within India’s constitutional order, statutory framework and judicial developments, and outlines a pragmatic regulatory pathway tailored to India’s socio-economic diversity.
BACKGROUND
OF THE STUDY India’s digital transformation (Digital India initiatives, large scale Aadhaar identity infrastructure and expanding internet penetration) has been accompanied by rapid adoption of data-driven and algorithmic systems in government and industry. State actors use automated tools for benefits targeting and surveillance; private platforms use recommendation algorithms, content moderation automation, targeted advertising and credit-scoring models. These systems
4 Justice K.S. Puttaswamy v. Union of India (2017)
5 Shreya Singhal v. Union of India (2015)
interact with vulnerabilities in governance capacity, socio-economic inequality, and legal gaps making legal accountability and socio-legal analysis urgent. Policy actors and think tanks (notably NITI Aayog’s National Strategy for AI) have advocated harnessing AI for inclusive growth, while legal scholarship and civil society flag risks to privacy, equality, and rule of law. The courts, most notably in the privacy jurisprudence culminating in Justice K. S. Puttaswamy, have created constitutional guardrails relevant to information governance and automated processing. Yet there is no consolidated Indian statute that deals specifically with algorithmic transparency, explainability, auditing, or liability for high-risk AI systems.
Existing instruments (Information Technology Act 2000; Digital Personal Data Protection Act 2023) provide partial tools but do not fully address algorithmic governance.
OBJECTIVE
OF THE STUDY Ø To map the statutory and judicial landscape in India relevant to AI accountability. Ø To identify socio-legal challenges arising from AI deployment across public and private spheres. Ø To analyse key case law principles bearing on automated decision-making, privacy and intermediary liability. Ø To propose a practical regulatory architecture and recommendations to ensure accountability, fairness, and redress in AI systems in India. LEGAL FRAMEWORKS IN INDIA WITH EXISTING LEGISLATION AND INSTRUMENTS India’s legal system is yet to develop a comprehensive legislative framework exclusively governing Artificial Intelligence (AI). However, several existing laws and policy instruments collectively regulate the indirect aspects of AI such as data protection, cybersecurity, intellectual property, contractual liability, and administrative accountability.
The absence of a dedicated “AI law” means that the current regulatory approach is fragmented and reactive rather than proactive. Nevertheless, the cumulative effect of these laws establishes the initial contours of accountability for AI-related activities within the Indian jurisdiction.
1. THE INFORMATION TECHNOLOGY ACT, 2000 (IT ACT) AND ITS
AMENDMENTS The Information Technology Act, 20006, remains the cornerstone of India’s cyber law infrastructure. Although it does not explicitly address AI, its provisions are invoked when AIdriven systems involve data misuse, cybercrime, or digital fraud. Section 43A mandates corporate entities to adopt “reasonable security practices” while handling sensitive personal data, thereby creating a basis for accountability in cases of AIbased data breaches or negligent data handling. Section 66 and Section 66E address unauthorized access and violation of privacy through electronic means, which can extend to AI-enabled surveillance or image recognition systems. Section 72A penalizes disclosure of information in breach of lawful contract, which becomes relevant where AI systems process or share personal information without user consent.
Judicial interpretation has expanded the IT Act’s applicability. In K. S. Puttaswamy v. Union of India 7(2017), the Supreme Court underscored informational privacy as a component of Article 21 of the Constitution8, emphasizing that digital technologies must respect data autonomy. AIdriven data processing must therefore align with constitutional and statutory obligations under the IT Act.
2. DIGITAL PERSONAL DATA PROTECTION ACT, 2023 (DPDP ACT)
The Digital Personal Data Protection Act, 20239, is India’s first comprehensive data protection legislation, inspired by global models like the GDPR. It directly impacts AI systems that depend on large-scale data collection and algorithmic profiling. Consent-Based Processing: AI developers must obtain explicit consent for data use, particularly for automated profiling or behavioural analysis. Right to Correction and Erasure: Individuals can demand correction of inaccurate AI-generated data, promoting algorithmic accountability. Data Fiduciary Obligations: AI-driven entities are treated as “data fiduciaries,” with duties to ensure fairness,
6 The Information Technology Act, 2000 Section 43A, Section 66, Section 66E, Section 72A
7 In K.S. Puttaswamy v. Union of India (2017)
8 Article 21 Constitution of india
9 The Digital Personal Data Protection Act, 2023
transparency, and proportionality. Cross-Border Data Transfers: The law restricts transfers of personal data outside India to ensure national data sovereignty a crucial aspect in AI cloud training models. This Act establishes the Data Protection Board of India, which serves as an adjudicatory body for violations, offering a quasi-judicial mechanism to ensure accountability of AI-based data handlers.
3. THE INDIAN PENAL CODE (IPC), 1860 / BHARATIYA NYAYA SANHITA, 2023
With the introduction of the Bharatiya Nyaya Sanhita (BNS), 202310, replacing the IPC, the criminal liability framework continues to apply to acts facilitated through AI systems. For instance, AI-generated misinformation, deepfakes, or cyberstalking can attract liability under provisions addressing defamation, fraud, or obscenity. Section 67 of the IT Act (read with relevant BNS provisions) penalizes publication of obscene material in electronic form, applicable to AI-generated content. Deepfake cases, such as those involving non-consensual morphing of women’s images, highlight the intersection between AI, dignity, and criminal accountability. Indian courts have begun to consider technological aspects in cybercrime adjudication, as seen in Sharat Babu Digumarti v.
Government of NCT of Delhi (2017)11, where the Supreme Court clarified that liability under the IT Act is distinct from traditional IPC offences.
4. CONSUMER PROTECTION ACT, 2019
The Consumer Protection Act, 201912, modernized India’s consumer law to include ecommerce and digital services. It allows users harmed by AI-based decision-making such as algorithmic mis-selling, biased recommendations, or defective AI services to seek remedies. Under Section 2(11), the definition of “deficiency in service” can extend to malfunctioning AI systems. The Consumer Protection (E-Commerce) Rules, 2020 also emphasize transparency and accountability in algorithmic advertising, thereby indirectly regulating AI-enabled commerce.
10 Bharatiya Nyaya Sanhita (BNS), 2023
11 Sharat Babu Digumarti v. Government of NCT of Delhi (2017)
12 The Consumer Protection Act, 2019 section 2(11)
Case precedents like Amazon Seller Services Pvt. Ltd. v. Amway India Enterprises Pvt. Ltd. (2019)13 highlight the judiciary’s approach towards digital accountability and platform responsibility, which can analogously apply to AI-driven intermediaries.
5. COPYRIGHT ACT, 1957 AND INTELLECTUAL PROPERTY LAWS14
AI’s ability to generate content has challenged traditional notions of authorship and originality. Under the Copyright Act, 195715, authorship is restricted to human creators. Therefore, AIgenerated works do not enjoy direct protection unless attributed to a human programmer or controller. In Tech Plus Media Pvt. Ltd. v. Jyoti Janda (2014)16, the Delhi High Court reaffirmed that originality requires human skill and judgment. This principal limit AI’s role as an author and raises questions about ownership and infringement liability. Similarly, the Patents Act, 197017, allows patenting of AI-assisted inventions but excludes autonomous AI creations, reflecting the absence of legal personality for AI entities.
6. COMPETITION ACT, 200218
The Competition Commission of India (CCI) plays a crucial role in regulating algorithmic collusion, predatory pricing, and data monopolies. AI systems used in market analysis or pricing algorithms may inadvertently facilitate anti-competitive practices. In Samir Agrawal v. ANI Technologies Pvt. Ltd. (2020)19, the CCI examined Uber and Ola’s algorithmic fare determination models, indicating judicial recognition of AI’s role in market conduct. The decision underscored the need for transparency and fairness in algorithmic decision-making within digital markets.
7. CONSTITUTIONAL AND ADMINISTRATIVE LAW DIMENSIONS
Beyond statutory provisions, India’s Constitution provides the fundamental normative basis for
13 Amazon Seller Services Pvt. Ltd. v. Amway India Enterprises Pvt. Ltd. (2019)
14 INTELLECTUAL PROPERTY LAWS
15 Copyright Act, 1957
16 In Tech Plus Media Pvt. Ltd. v. Jyoti Janda (2014)
17 The Patents Act, 1970
18 COMPETITION ACT, 2002
19 Samir Agrawal v. ANI Technologies Pvt. Ltd. (2020)
AI regulation. Article 14 20(Equality before Law) mandates fairness and non-discrimination, which applies to algorithmic decision-making. Article 19(1)(a) 21guarantees freedom of expression, relevant to AI-generated content moderation. Article 21 ensures the right to life and privacy, as articulated in Puttaswamy (2017) and later reaffirmed in Internet and Mobile Association of India v. RBI (2020).22 Administrative law principles such as natural justice, reasoned decisions, and accountability— apply to AI use in governance. The Delhi Police’s use of facial recognition or the Income Tax Department’s AI-based profiling must comply with procedural fairness under Article 14.
8. NITI AAYOG’S POLICY INSTRUMENTS
In the absence of a specific law, the Government’s policy documents serve as guiding instruments: NITI Aayog’s “National Strategy for Artificial Intelligence (AI for All)” (2018) emphasized ethical, inclusive, and transparent AI deployment. Responsible AI for All: 2021 Approach Document proposed accountability, fairness, and explainability principles. Digital India Act (Proposed, 2024) aims to replace the IT Act with a broader digital governance law, incorporating AI-specific ethical guidelines and liability mechanisms. These instruments reflect India’s intent to balance innovation with responsible regulation.
19. SECTOR-SPECIFIC REGULATIONS
Certain sectors already implement AI-related regulatory measures RBI’s Guidelines on Digital Lending (2022) 23address algorithmic credit scoring and ensure human oversight. IRDAI’s Sandbox Regulations encourage ethical AI use in insurance
20 Article 14 constitution of India
21 Constitution Of India Article 19(1)(a) guarantees freedom of expression, relevant to AI-generated content moderation.
22 Mobile Association of India v. RBI (2020).
23 RBI’s Guidelines on Digital Lending (2022)
underwriting. SEBI’s Algorithmic Trading Guidelines monitor AI-driven securities transactions. Telecom Regulatory Authority of India (TRAI) has discussed AI use in content moderation and spam regulation. Together, these frameworks create a mosaic of sectoral accountability even without a unified AI statute.
CONTEMPORARY LEGAL AND SOCIO-LEGAL
CHALLENGES
WITH CASE LAW Artificial Intelligence (AI) has swiftly permeated India’s legal, administrative, and commercial frameworks, posing intricate socio-legal challenges that span privacy, discrimination, liability, and constitutional governance. The integration of AI-driven systems in both public and private domains has redefined traditional notions of accountability and justice. From automated content moderation to predictive policing, algorithmic decision-making challenges the very foundations of the rule of law — transparency, fairness, and reasoned decision-making. This section examines key contemporary challenges in India, anchored in relevant jurisprudence.
1. Privacy, Surveillance, and Data Protection
The foremost legal challenge surrounding AI in India is the tension between innovation and privacy. AI thrives on vast datasets — often personal or sensitive — to train algorithms and generate insights. However, unchecked data collection and profiling violate individuals’ right to privacy, dignity, and informational autonomy. The landmark decision in Justice K. S. Puttaswamy (Retd. ) v. Union of India (2017) 24established privacy as a fundamental right under Article 21 of the Constitution, declaring informational self-determination an integral part of personal liberty. The subsequent Puttaswamy (Aadhaar Case) (2018) 25reiterated that any intrusion into privacy must satisfy the test of legality, necessity, and proportionality. Yet, state-led AI systems — such as facial recognition in policing and welfare targeting — often function without explicit statutory authorization.
Surveillance systems deployed by police departments in Telangana and Delhi for “crime mapping” and “crowd analytics” operate in a legal vacuum, conflicting with the proportionality test laid down in Anuradha Bhasin v. Union of India (2020)26, which underscored procedural safeguards against arbitrary state action in the digital sphere.
24 The landmark decision in Justice K.S. Puttaswamy (Retd.) v. Union of India (2017)
25 Puttaswamy (Aadhaar Case) (2018)
26 Anuradha Bhasin v. Union of India (2020)
The Digital Personal Data Protection Act, 202327, though a step forward, lacks specific provisions regulating automated decision-making, algorithmic profiling, or explainability obligations. Hence, individuals affected by automated decisions have limited grounds for redress, unlike the “right to explanation” enshrined under the EU’s General Data Protection Regulation (GDPR).
2. Algorithmic Bias and Discrimination
AI models are only as unbiased as the data used to train them. In India’s stratified social structure, data sets often reflect existing inequalities of caste, gender, and class. Algorithmic bias thus perpetuates systemic discrimination in credit scoring, hiring, predictive policing, and welfare delivery. The constitutional principle of equality enshrined in Articles 14 and 15 mandates nondiscrimination and fairness. The Supreme Court’s decisions in State of West Bengal v. Anwar Ali Sarkar (1952) and E. P. Royappa v. State of Tamil Nadu (1974) 28extended equality to prohibit arbitrariness. These principles, applied to algorithmic contexts, imply that AI systems used by public authorities must satisfy reasonableness and non-arbitrariness. Similarly, in Navtej Singh Johar v.
Union of India (2018)29, the Court emphasized dignity and equality in digital and physical spheres alike — a benchmark for assessing algorithmic discrimination. The lack of transparency in AI models often makes proving such bias difficult. The absence of standardized auditing or documentation (e. g. , “model cards” or “data sheets”) further compounds evidentiary challenges. From a socio-legal perspective, algorithmic bias has deeper implications: it amplifies castebased targeting in predictive policing or welfare exclusion. Marginalized communities, already under-represented in datasets, are systematically disadvantaged — reinforcing inequality under the guise of neutrality.
3. Transparency and Explainability
AI’s opacity often termed the “black box problem” undermines procedural fairness and
27 The Digital Personal Data Protection Act, 2023
28 Ali Sarkar (1952) and E.P. Royappa v. State of Tamil Nadu (1974)
29 Navtej Singh Johar v. Union of India (2018)
accountability. When government agencies rely on AI for decision-making (such as in tax risk scoring or welfare eligibility), individuals are entitled to know how and why a decision was reached. In Maneka Gandhi v. Union of India (1978), 30the Court held that procedure established by law must be “fair, just, and reasonable. ” This principle extends naturally to algorithmic governance: automated decisions without explanation violate due process. The absence of a right to explanation in Indian law creates a vacuum. While the Information Technology Act, 2000 regulates data intermediaries, it does not impose algorithmic transparency requirements. The Delhi High Court in K. S. Puttaswamy (Aadhaar)31 highlighted the necessity of “meaningful consent,” which is rendered illusory in automated systems that operate opaquely. Globally, explainability has become a cornerstone of AI regulation (e. g.
, the EU’s AI Act). India’s lack of equivalent obligations risks deepening the accountability gap between algorithmic governance and constitutional protections.
4. Freedom of Speech, Expression, and Automated Censorship
AI-powered moderation tools on social media platforms and automated takedown mechanisms pose critical challenges to freedom of expression. In Shreya Singhal v. Union of India (2015)8, the Supreme Court struck down Section 66A of the IT Act, holding that vague and arbitrary restrictions on online speech violated Article 19(1)(a). However, the evolution of automated moderation systems — often operating without human review — risks reintroducing censorship through technology. Intermediary liability jurisprudence, beginning with Avnish Bajaj v. State (NCT of Delhi) (2005) 32and continuing through MySpace Inc. v. Super Cassettes Industries Ltd. (2016)33, established conditional immunity for platforms acting in “good faith. ” Yet, when automated filters misclassify content, users are deprived of expressive freedom without procedural recourse.
The new IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, mandate prompt removal of “unlawful” content, incentivizing overbroad algorithmic censorship.
30 In Maneka Gandhi v. Union of India (1978),
31 The Delhi High Court in K.S. Puttaswamy (Aadhaar)
32 Avnish Bajaj v. State (NCT of Delhi) (2005)
33 MySpace Inc. v. Super Cassettes Industries Ltd. (2016)
This automation of moderation absent human oversight clashes with S. Rangarajan v. P. Jagjivan Ram (1989), which underscored that freedom of expression cannot be curtailed based on speculative harm. A socio-legal reading suggests that algorithmic governance, when applied to expression, must preserve both procedural fairness and contextual sensitivity capacities that AI lacks.
5. State Surveillance, Predictive Policing, and Civil Liberties
Predictive policing and facial recognition deployments represent one of the gravest AI-related threats to civil liberties in India. Police departments in Delhi and Hyderabad use facial recognition for crowd control and criminal identification. These systems often lack statutory authorization or independent oversight. In People’s Union for Civil Liberties v. Union of India (1997)34, the Supreme Court mandated procedural safeguards for telephone interception, recognizing privacy as implicit in Article 21. Extending that logic, mass biometric surveillance without statutory basis violates constitutional principles of legality and proportionality reaffirmed in Puttaswamy and Anuradha Bhasin. Globally, courts have curtailed similar practices: in Bridges v. South Wales Police (UK, 2020)35, facial recognition was held to breach privacy rights.
India lacks comparable jurisprudence but must anticipate similar concerns. Predictive policing based on flawed datasets risks perpetuating caste or communal profiling, contravening Article 15’s prohibition of discrimination and Article 19(1)(d)’s guarantee of free movement.
6. Liability and Accountability for AI-Driven Harm
When AI systems cause harm a wrongful credit denial, medical misdiagnosis, or vehicular accident determining liability becomes complex. Indian tort and consumer laws require human fault, but AI complicates causation. The Consumer Protection Act, 2019 allows redress for defective products and services, yet fails to classify algorithmic systems as products with legal personhood.
34 Union for Civil Liberties v. Union of India (1997)
35 Bridges v. South Wales Police (UK, 2020)
In Donoghue v. Stevenson (1932, UK)36, the “neighbour principle” laid the foundation for modern negligence law. Indian courts have expanded this doctrine, as seen in Municipal Corporation of Delhi v. Subhagwanti (1966)37, imposing strict liability for hazardous activities. Extending this reasoning, developers and deployers of high-risk AI should bear strict or vicarious liability for algorithmic harms. The Indian Medical Association v. V.P. Shantha (1995)38 judgment, which recognized medical negligence under consumer law, can analogically apply to AI diagnostic tools: doctors or institutions relying solely on AI without due diligence could be liable for negligence.
7. Evidentiary Challenges and Digital Proof
AI-generated evidence such as predictive analytics, surveillance data, or algorithmic outputs introduces novel evidentiary questions. The Indian Evidence Act, 1872 (as amended by Section 65B of the IT Act) provides for electronic records but does not contemplate algorithmic outputs. In Anvar P.V. v. P.K. Basheer (2014)39, the Supreme Court emphasized strict compliance with Section 65B certification for electronic evidence. AI systems’ opacity complicates authentication and cross-examination. Courts face difficulty in verifying algorithmic reasoning, leading to potential miscarriages of justice. State of Maharashtra v. Bharat Shantilal Shah (2008) further highlighted the need for stringent oversight in digital interception and data handling. Hence, evidentiary standards must evolve to accommodate AI’s probabilistic nature while safeguarding procedural fairness.
8. Socio-Economic Exclusion and the Digital Divide
Beyond doctrinal issues, AI’s socio-legal impact manifests in structural inequities. Algorithmic governance disproportionately affects marginalized groups who lack digital literacy, documentation, or access to grievance mechanisms. The Aadhaar litigation (2018) demonstrated how biometric mismatches led to denial of welfare entitlements, raising due process concerns under Article 21.
36 In Donoghue v. Stevenson (1932, UK)
37 Municipal Corporation of Delhi v. Subhagwanti (1966)
38 The Indian Medical Association v. V.P. Shantha (1995)
39 In Anvar P.V. v. P.K. Basheer (2014)
Similarly, exclusion from automated recruitment or loan-assessment systems perpetuates economic marginalization. In Olga Tellis v. Bombay Municipal Corporation (1985)40, the Court held that the right to livelihood is inherent in the right to life. Extending this to AI contexts, unjust algorithmic exclusion violates constitutional guarantees of livelihood and equality. AI thus becomes a mirror reflecting existing social hierarchies, necessitating socio-legal interventions that emphasize inclusivity, participatory design, and digital literacy.
9. Intellectual Property and Data Ownership
AI raises complex intellectual property questions: who owns AI-generated works, and is using copyrighted data for model training permissible? While Indian copyright law (Copyright Act,
1957) protects works of human authorship, it remains silent on machine-generated outputs.
The Eastern Book Company v. D.B. Modak (2008) 41case clarified the “modicum of creativity” test for originality. Applied to AI, outputs generated autonomously might not satisfy this threshold. The absence of explicit AI authorship provisions creates uncertainty for creators, developers, and rights holders. Data ownership, another grey area, intersects with privacy and contractual obligations. AI’s dependence on data aggregation highlights the urgent need for clarifying data proprietorship within India’s emerging digital economy.
10. Institutional Capacity and Regulatory Fragmentation
India’s regulatory landscape for AI is fragmented across ministries and sectoral regulators, NITI Aayog, RBI, IRDAI, SEBI, and the Ministry of Health. The absence of a central AI regulatory authority leads to overlap and inconsistency. The Supreme Court in Vineet Narain v. Union of India (1998)42 underscored the necessity of independent regulatory bodies insulated from executive influence a principle equally applicable to AI oversight. Effective AI regulation requires techno-legal expertise, multi-stakeholder participation, and cross-sector coordination. Without institutional capacity, even progressive legislation will fail
40 In Olga Tellis v. Bombay Municipal Corporation (1985)
41 The Eastern Book Company v. D.B. Modak (2008)
42 Vineet Narain v. Union of India (1998)
in implementation, perpetuating regulatory capture and technological opacity.
11. Comparative Global Insights
While India’s courts are yet to directly adjudicate AI liability, comparative jurisprudence offers guidance. The EU’s AI Act (2024 draft) introduces a risk-based classification, mandating stringent obligations for “high-risk AI.” The US Algorithmic Accountability Act proposal emphasizes transparency and auditing. India can adapt these principles within its constitutional and socio-economic context, ensuring alignment with the doctrines of proportionality and fundamental rights protection articulated in Puttaswamy.
12. The Socio-Legal Imperative
AI accountability cannot be confined to technical compliance or statutory regulation alone. The socio-legal perspective demands embedding ethics, human rights, and distributive justice within technological governance. Law must evolve not merely as a reactive mechanism but as a proactive framework ensuring equitable access and protection. Courts, regulators, and civil society must collaborate to ensure that technological progress does not erode constitutional morality a principle reaffirmed in K.S. Puttaswamy, Navtej Johar, and Joseph Shine v. Union of India (2018)43, which collectively expanded the understanding of dignity and autonomy. Thus, India stands at a pivotal moment: the legal system must move from reactive adjudication to anticipatory governance, embedding accountability in the very design of AI systems.
SUGGESTIONS AND
RECOMMENDATIONS
India stands at a pivotal juncture in shaping a comprehensive legal and ethical ecosystem for Artificial Intelligence (AI). The existing frameworks the Information Technology Act, 2000 and the Digital Personal Data Protection Act, 2023 offer only fragmented protections. A coordinated and forward-looking strategy is essential to ensure that AI deployment remains consistent with constitutional values of liberty, equality, and dignity. The following recommendations aim to strengthen the socio-legal accountability of AI in India.
43 Joseph Shine v. Union of India (2018)
1. ENACT A DEDICATED “ARTIFICIAL INTELLIGENCE ACCOUNTABILITY AND
GOVERNANCE ACT.”
India requires a unified legislative framework that defines key concepts such as “automated decision-making,” “algorithmic bias,” and “high-risk AI systems.” The statute should establish obligations on developers and deployers to conduct algorithmic impact assessments before deployment, ensure explainability, and provide mechanisms for redress in case of harm. The Act should also mandate proportional penalties for non-compliance, modelled on risk-based approaches used in the EU’s AI Act.
2. CREATE AN INDEPENDENT AI REGULATORY AUTHORITY.
A specialized body preferably under the Ministry of Electronics and Information Technology should be empowered to monitor, audit, and certify AI systems across sectors. This authority must collaborate with existing regulators (RBI, IRDAI, TRAI, and SEBI) to harmonize sectoral standards and prevent regulatory overlap. It should maintain a public registry of high-risk AI systems and require regular algorithmic audits by accredited third-party experts.
3. STRENGTHEN DATA PROTECTION AND USER RIGHTS.
The DPDP Act should be expanded to explicitly address automated profiling and the right to explanation. Individuals must have the legal right to opt out of fully automated decisionmaking where significant effects on their rights or livelihoods are involved. Transparent consent architecture and stronger cross-border data safeguards are vital to mitigate misuse of Indian citizens’ data in global AI models.
4. ENSURE HUMAN OVERSIGHT AND ACCOUNTABILITY.
Automated decision-making must always include a human-in-the-loop in critical areas such as law enforcement, healthcare, and welfare delivery. Government use of AI for surveillance or predictive policing should be subject to judicial authorization and legislative oversight. Administrative law principles of fairness, reasonableness, and proportionality must guide all algorithmic applications affecting citizens.
5. PROMOTE ALGORITHMIC TRANSPARENCY AND FAIRNESS AUDITS.
All public and private entities deploying AI should publish “Model Cards” disclosing data sources, limitations, bias-mitigation methods, and validation results. Mandatory fairness and non-discrimination audits will enhance public trust and enable regulatory enforcement against algorithmic bias.
6. BUILD INSTITUTIONAL AND JUDICIAL CAPACITY.
Judges, bureaucrats, and law-enforcement officials should receive continuous training in digital evidence, data analytics, and algorithmic governance. Law schools and research institutions should be encouraged to develop specialized courses on AI law and ethics to produce technolegal professionals capable of bridging the gap between law and technology.
7. ENCOURAGE RESPONSIBLE INNOVATION AND PUBLIC AWARENESS.
Government should incentivize ethical AI start-ups and open-source tools promoting inclusivity, multilingual accessibility, and environmental sustainability. Public awareness campaigns can help citizens understand data rights and risks associated with AI.
8. FOSTER INTERNATIONAL COOPERATION.
India must engage actively in global AI governance forums to shape standards that reflect its constitutional and developmental priorities. Bilateral and multilateral collaborations can ensure interoperability of regulations and facilitate responsible cross-border data flows.
CONCLUSION
Artificial Intelligence (AI) represents both a transformative opportunity and a profound regulatory challenge for India’s legal and social order. As AI systems increasingly influence governance, commerce, and citizen welfare, ensuring accountability, transparency, and human rights protection becomes imperative. The present Indian legal architecture—comprising the Information Technology Act, 2000, and the Digital Personal Data Protection Act, 2023—offers a foundational framework but lacks the precision to address complex issues such as algorithmic bias, automated decision-making, and ethical accountability. The absence of a unified statutory regime leaves a regulatory vacuum that risks undermining constitutional guarantees of equality, privacy, and due process. Judicial interventions, particularly in Justice K. S. Puttaswamy v. Union of India and Shreya Singhal v.
Union of India, have affirmed privacy and freedom of expression as essential safeguards against technological overreach. Yet, as AI applications expand to domains like surveillance, predictive policing, and digital governance, mere reliance on judicial oversight is inadequate. A comprehensive legal architecture must integrate algorithmic auditing, human oversight, and institutional mechanisms capable of monitoring high-risk AI systems. Socio-legally, AI governance in India must reflect principles of inclusivity and distributive justice. The use of AI in welfare schemes, credit scoring, and employment decisions must not entrench existing inequalities or marginalize vulnerable communities. Legislative reforms, capacity-building measures, and public participation in AI policy-making are vital to bridging the gap between innovation and ethics.
In conclusion, India’s path forward lies in adopting a rights-based, risk-sensitive, and transparent approach to AI governance. A dedicated “Artificial Intelligence Accountability and Governance Act,” coupled with an empowered regulatory authority and sectoral harmonization, can anchor AI development within the constitutional ethos of justice, liberty, and equality. Such a framework will ensure that India’s AI revolution evolves not as a tool of control, but as a catalyst for inclusive and responsible progress.
REFERENCES
Statutes and Government Instruments The Information Technology Act, 2000 (Act No. 21 of 2000). The Digital Personal Data Protection Act, 2023 (Act No. 22 of 2023). The Indian Penal Code, 1860. The Constitution of India, 1950. The Aadhaar (Targeted Delivery of Financial and Other Subsidies, Benefits and Services) Act, 2016. The Right to Information Act, 2005. The Indian Evidence Act, 1872 (as amended). Draft National Strategy for Artificial Intelligence – “#AIforAll”, NITI Aayog, 2018. Ministry of Electronics and Information Technology (MeitY), “Responsible AI for All: Operationalising Principles for India” (Report, 2021). European Union, EU Artificial Intelligence Act (2024), for comparative analysis.
CASE LAW
REFERENCES
Justice K.S. Puttaswamy (Retd.) v. Union of India, (2017) 10 SCC 1. Shreya Singhal v. Union of India, (2015) 5 SCC 1. Anuradha Bhasin v. Union of India, (2020) 3 SCC 637. State of Maharashtra v. Praful B. Desai, (2003) 4 SCC 601. People’s Union for Civil Liberties v. Union of India, (1997) 1 SCC 301. Internet and Mobile Association of India v. Reserve Bank of India, (2020) 10 SCC 274.
Kharak Singh v. State of U. P. , AIR 1963 SC 1295. Naz Foundation v. Government of NCT of Delhi, (2009) 160 DLT 277. Maneka Gandhi v. Union of India, (1978) 1 SCC 248. Ritesh Sinha v. State of Uttar Pradesh, (2019) 8 SCC 1. Avnish Bajaj v. State (NCT of Delhi), (2005) 3 Comp LJ 364 (Del). Yahoo India Pvt. Ltd. v. Union of India, (2019) 13 SCC 232. Tata Press Ltd. v. Mahanagar Telephone Nigam Ltd. , (1995) 5 SCC 139. PUCL v. Union of India (Telephone Tapping Case), (1997) 1 SCC 301. K. S. Puttaswamy (Aadhaar-5J) v. Union of India, (2019) 1 SCC 1. Google India Pvt. Ltd. v. Visaka Industries Ltd. , (2020) 4 SCC 162. Bachan Singh v. State of Punjab, (1980) 2 SCC 684 — for proportionality principle. State of West Bengal v. Committee for Protection of Democratic Rights, (2010) 3 SCC 571. Navtej Singh Johar v. Union of India, (2018) 10 SCC 1. Modern Dental College & Research Centre v.
State of Madhya Pradesh, (2016) 7 SCC 353. BOOKS AND JOURNAL ARTICLES Sharma, R. (2022). Artificial Intelligence and Law: Challenges and Prospects in India. Eastern Book Company, Lucknow. S. Baxi and P. Singh, “AI and Fundamental Rights: A Constitutional Perspective,” (2021) 63 Journal of the Indian Law Institute 115. Singh, V. & Gupta, N. (2020). “Algorithmic Bias and Legal Responsibility: Emerging Concerns in Indian Jurisprudence. ” NUJS Law Review, Vol. 13, pp. 87–110.
Rajiv Kumar & A. Mehta, “Data Protection and Artificial Intelligence in India: Towards a Unified Legal Framework,” (2023) Indian Journal of Law and Technology Vol. 19(2), 201– 227. Mehta, P. (2021). Digital Constitutionalism in India: Rights and Technology in the 21st Century. Oxford University Press. UNESCO, “Ethics of Artificial Intelligence: Recommendation Adopted by the General Conference” (2021). OECD, “OECD Principles on Artificial Intelligence” (2019). C. Cath et al. , “Artificial Intelligence and the ‘Good Society’: The US, EU, and UK Approaches” (2018) 24 Science, Technology and Society 335. Kuner, C. (2020). “Data Sovereignty and AI Regulation: Comparative Global Developments,” International Data Privacy Law Vol. 10, pp. 89–102.
Srikrishna Committee Report, “A Free and Fair Digital Economy: Protecting Privacy, Empowering Indians” (2018), Ministry of Electronics and Information Technology. REPORTS AND POLICY DOCUMENTS NITI Aayog, National Strategy for Artificial Intelligence: #AIforAll (2018). Ministry of Electronics and IT, India AI Report: Responsible AI for Social Empowerment (RAISE) (2020).
World Economic Forum, Artificial Intelligence Governance Framework (2022). Internet Freedom Foundation, Algorithmic Accountability in India (Policy Brief, 2023). Reserve Bank of India, Report on FinTech and Digital Governance (2021).