Open Access Research Article
White Black Legal – International Law Journal · ISSN 2581-8503
CONSTITUTIONAL CHALLENGES OF AI-BASED SURVEILLANCE IN INDIA
Read the Full Research Paper
Access the complete open-access article in PDF format. No login is required.
Abstract
CONSTITUTIONAL CHALLENGES OF AI-BASED
SURVEILLANCE IN INDIA
Examining Facial Recognition Technology, Privacy,
Judicial Precedents and Regulatory Needs
AUTHORED
BY - ROHAN V, KARTHIKEYAN S & BHARATH S
Abstract
Artificial intelligence
is changing the way governments and law-enforcement agencies identify people,
manage information and conduct surveillance. Facial Recognition Technology
(FRT) is one of the most significant examples because it can compare facial images
with stored records and, in some settings, operate at scale. The technology may
assist legitimate functions such as identifying suspects or missing persons,
but its use also raises questions about privacy, equality, personal liberty and
democratic freedoms. This paper examines the constitutional issues surrounding
AI-based surveillance in India, with particular attention to Articles 14, 19
and 21 of the Constitution. It
studies the development of privacy jurisprudence and considers the relevance of
decisions including Justice K.S.
Puttaswamy (Retd.) v. Union of India, Maneka
Gandhi v. Union of India, Anuradha Bhasin v. Union of India and
the Delhi High Court proceedings in Sadhan
Haldar v. State (NCT of Delhi). The paper also considers the Criminal Procedure
(Identification) Act, 2022 and the Digital Personal Data Protection Act, 2023
as parts of the wider legal environment, while recognising that neither
provides a complete, technology-specific framework for every form of FRT
deployment. The paper argues for a rights-centred approach in which security
objectives are pursued through clear legal authority, defined purposes,
proportionality, data safeguards, accuracy testing, human review and meaningful
accountability. The aim is not to treat technology and constitutional rights as
opposing ideas, but to identify legal conditions under which innovation can
operate consistently with privacy, equality, liberty and the rule of law.
Keywords:
Artificial
Intelligence; Facial Recognition Technology; Surveillance; Privacy; Article 14;
Article 19; Article 21; Proportionality; Algorithmic Bias; Data Protection.
1.
Introduction
Artificial intelligence
has moved from being a specialised technological field to becoming part of
everyday governance and public administration. One of its most visible
applications is Facial Recognition Technology (FRT), which uses computer vision
and machine-learning techniques to identify or verify people from facial
images. When connected to cameras and databases, the technology can make
identification much faster and can also operate on a large scale.
In India, the use of
facial recognition has developed alongside broader digital governance, policing
and biometric identification
initiatives. The technology has attracted attention because it can serve
practical purposes, including the
identification of persons of interest and efforts to locate missing persons. At
the same time, the same technology can be used to observe and identify people
who are not suspected of wrongdoing. That difference matters constitutionally.
The central question is
therefore not whether FRT is useful. The more difficult question is how the
State should exercise this capability without
allowing surveillance to become arbitrary, excessive or insufficiently accountable.
The Supreme
Court's recognition of privacy as a fundamental right has made this question
particularly important.
Any analysis of AI-based surveillance must consequently consider the relationship between technological
capability and constitutional limits.
This paper focuses on
FRT as a concrete example of AI-based surveillance. It examines the
implications for equality under Article 14, freedoms under
Article 19 and privacy and personal liberty
under Article 21. It then considers
judicial principles, existing statutory provisions
and the need for additional
safeguards.
2.
Statement
of the Problem
The rapid development
of facial recognition creates a gap between what technology makes possible and
what the law clearly regulates. A camera system can record a face, an algorithm
can convert the image into a
biometric template, and a database can be searched for a possible match. Each
stage raises a different legal question: who may collect the data, for what
purpose, under what authority, how long it may be retained, who may access it,
and what happens when the system is wrong?
The problem
becomes more serious
when a system moves from limited identification to continuous or widespread
monitoring. In such circumstances, individuals may have little practical knowledge
of the processing taking place.
They may also have
difficulty challenging an incorrect match or finding out how an automated
decision was reached.
The research
therefore addresses the constitutional and regulatory uncertainty surrounding the use of
FRT by public authorities. It does not assume that every use of facial
recognition is unlawful. Instead, it asks what legal conditions and safeguards
are required for such use to remain compatible with fundamental rights.
3.
Objectives
of the Study
The main objective of this research
is to examine the constitutional challenges created by AI-based
surveillance, with FRT as the principal case study.
The specific objectives are:
1. To examine the relationship between FRT and Articles 14, 19 and 21 of the Constitution.
2. To explain the relevance
of the constitutional right to privacy to biometric surveillance.
3. To analyse judicial principles that regulate State interference with fundamental rights.
4. To examine
relevant statutory developments, including the Criminal
Procedure (Identification) Act, 2022 and the Digital Personal Data Protection
Act, 2023.
5. To identify
concerns relating to mass surveillance, function creep, algorithmic bias and wrongful identification.
6. To suggest practical legal and institutional safeguards for responsible deployment.
4.
Research
Questions
The study is guided by the following questions:
1. What constitutional rights are most directly affected
by AI-based facial surveillance?
2. How does the right to privacy limit State use of FRT?
3. What Article 14 concerns can arise from inaccurate
or discriminatory algorithmic outcomes?
4. Can widespread
facial tracking affect freedoms protected
by Article 19?
5. Do existing
Indian laws provide
sufficient safeguards for every form of FRT deployment?
6. What
regulatory and institutional measures can improve transparency, accountability
and protection of individual rights?
5.
Research
Methodology
This is a doctrinal and
analytical legal study. It examines constitutional provisions, judicial
decisions, legislation and policy material relating
to privacy, surveillance, biometric identification and AI-based
systems.
The research uses
primary legal sources to identify constitutional principles and secondary
literature to understand the technological and policy concerns
surrounding FRT.
The study is qualitative rather than empirical.
It does not claim to measure the accuracy of a particular facial recognition product or to determine the frequency of
wrongful matches in India. Instead, it examines the legal questions that arise from the capability of FRT and assesses those questions against
constitutional standards.
The analysis is
deliberately framed around the distinction between legitimate, targeted uses
and broad or continuous surveillance.
This distinction is important because the constitutional assessment depends not
only on the technology itself but also on the purpose,
scale, safeguards and consequences of its use.
6.
Understanding
Facial Recognition Technology
Facial Recognition Technology is a biometric technology that uses facial characteristics to identify or verify a person. A simplified FRT process begins
with image capture.
Software then detects
a face, identifies relevant facial features, converts
them into a mathematical representation and compares that representation with images
or templates in a database.
The technology can be used for one-to-one verification, where a person claims an identity,
or for one-to-many identification,
where a face is compared against a database to find possible matches. The
second form creates more significant surveillance concerns because a person can be identified without actively presenting an identity claim.
FRT therefore has two sides.
On one side, it can support investigations and help authorities work with large
amounts of visual information. On the other, its scale can reduce practical
anonymity in public spaces and create consequences
when a system produces a false match.
Legal regulation must address both sides rather
than treating FRT as either inherently beneficial
or inherently harmful.
7.
Constitutional
Framework: Articles 14, 19 and 21
7.1 Article 14 – Equality
and Non-Arbitrariness
Article 14 requires the State to act consistently with equality before the law and equal protection of the
laws.
In the context of FRT, concerns
can arise where
an algorithm produces
systematically different error rates across groups, or where authorities use the technology in a selective or arbitrary manner.
A false match can also
have serious consequences if
it becomes the basis for further police action. These concerns make accuracy, testing,
record-keeping and review important safeguards.
7.2 Article 19 – Speech and Peaceful
Assembly
Article 19(1)(a)
protects freedom of speech and expression, while
Article 19(1)(b) protects
the right to assemble
peaceably and without
arms, subject to constitutional restrictions. Pervasive identification at public gatherings
may create a chilling effect if people reasonably believe
that participation will be individually recorded and identified.
The legal question is therefore not simply whether surveillance exists, but
whether its purpose, method and scale are justified by law.
7.3 Article 21 – Privacy and Personal
Liberty
Article 21 is central
to the legal analysis because facial recognition involves the processing of
information linked to an identifiable individual. The Supreme Court's
privacy jurisprudence recognises privacy as part of
constitutional liberty and dignity.
The collection and use of facial information therefore require careful
consideration of legality, purpose and proportionality.
8.
Privacy
and the Proportionality Framework
The nine-judge bench decision in Justice K.S. Puttaswamy (Retd.) v. Union of India (2017) recognised the right to privacy as a fundamental right. The judgment placed
privacy within the constitutional protection of Part III and connected
it with liberty, dignity and autonomy.
The proportionality analysis developed in subsequent constitutional jurisprudence is particularly useful forsurveillance questions. A rights-intrusive measure
should have a lawful basis
and pursue a legitimate
objective.
There should
be a rational connection between
the measure and that objective, and the State
should consider whethera less restrictive alternative is available. Finally, the impact on rights must be balanced against
the importance of the public objective.
For FRT, this means
that the existence of a security objective does not end the constitutional
inquiry. The authority using the system should be able to identify
the legal basis, the specific
purpose, the categories of persons affected, the safeguards in place and the reasons
why the chosen level of surveillance is necessary. This approach
also helps distinguish targeted identification from indiscriminate monitoring.
9.
Judicial
Precedents Relevant to AI Surveillance
9.1 Justice K.S. Puttaswamy (Retd.)
v. Union of India
The 2017 privacy judgment
is the constitutional starting point for analysing
modern surveillance. It confirms that privacy is a fundamental
right and provides the foundation for examining State action that interferes
with privacy.
9.2 Maneka Gandhi v. Union of India
The judgment
gave Article 21 a broad
and substantive interpretation and emphasised that procedures
affecting personal liberty must satisfy constitutional requirements of fairness
and reasonableness. Its wider significance is
that liberty cannot be separated
from the quality of the procedure used by the State.
9.3 Anuradha Bhasin v. Union of India
The Supreme
Court applied proportionality principles while considering restrictions on fundamental freedoms. The case is relevant
to surveillance because
it demonstrates the need to examine the justification, scope
and duration of rights-restrictive measures.
9.4 Sadhan Haldar v. State (NCT of Delhi)
The Delhi High Court
proceedings are relevant to the use of facial recognition in the context of
missing children.
The orders record discussion of the Delhi Police's facial-recognition software and the need to improve
mechanisms for locating missing children. The case is better understood as an
example of an identified and limited use context;
it should not be cited as a general
Supreme Court ruling declaring that FRT is constitutionally valid or
invalid.
9.5 PUCL v. Union of India
The telephone-tapping
jurisprudence of PUCL is important to the broader law of surveillance because
it illustrates the constitutional importance of procedural safeguards when the State exercises
intrusive monitoring powers.
10.
Statutory
and Regulatory Landscape
India's legal framework relevant
to biometric and digital surveillance is spread across different
instruments. This fragmented structure makes it important to examine each law
for its actual scope rather than treating any one statute as a complete FRT law.
The Criminal Procedure
(Identification) Act, 2022 expands the kinds of measurements that may be taken
for specified criminal-justice purposes and provides for collection, storage
and sharing of records. It is relevant
to biometric identification, but it does not amount to a comprehensive statute
regulating every public-space facial recognition deployment.
The Digital
Personal Data Protection Act, 2023 also forms part of the data-governance framework.
Section 17 contains specific exemptions, including provisions concerning processing for prevention, detection, investigation or prosecution
of offences and certain notified State instrumentalities. These exemptions make
it especially important to examine how surveillance practices are governed
by other legal rules and constitutional standards.
The important point is
that the existence of data-protection legislation does not, by itself, resolve
every constitutional question concerning facial recognition. Issues of statutory
authority, purpose, proportionality, accuracy, oversight
and remedies continue
to require separate
consideration.
11.
Constitutional
Challenges in FRT Deployment
11.1 Mass or Indiscriminate Surveillance
The greatest
constitutional concern arises
when facial recognition changes from targeted
identification into continuous monitoring. A system that attempts to identify everyone
passing a camera is materially different from a system used to verify a person in a defined investigation.
11.2 Function Creep
Information collected
for one purpose may later be used for another.
This can undermine
purpose limitation and make it difficult for individuals to know how their biometric
information is being used. Purpose
limitation is therefore a practical constitutional safeguard as well as a data-governance principle.
11.3 Algorithmic Bias
Facial recognition systems can produce
errors, and performance can vary depending on the system, dataset and conditions in which
the technology is used. Where errors have unequal effects, Article 14 concerns
may arise.
Independent testing and transparent performance reporting are consequently important.
11.4 Wrongful Identification
A facial
match is an investigative lead,
not necessarily proof
of identity or guilt. Treating
an automated match as conclusive can create serious
risks of mistaken
police action. Human verification and procedural
safeguards are necessary before a match produces adverse consequences.
11.5 Chilling Effects
People may behave
differently when they believe that their movements or participation in public
activities can be identified and recorded.
This concern connects
privacy with the practical exercise
of freedoms under Article 19.
12.
Regulatory
Gaps and Accountability
The main regulatory challenge is not simply
that technology is developing quickly.
It is that the legal framework
must clearly allocate
responsibility at every stage of the surveillance process.
A robust system should
identify the authority responsible for deployment, define permissible purposes, establish retention periods, control access
to biometric databases and record how searches are conducted. It should also provide independent mechanisms for auditing
accuracy and investigating misuse.
Transparency is equally important. Individuals cannot effectively challenge a system
if the government agency does not disclose basic information about its purpose,
legal authority and safeguards. At the same time,
legitimate security interests
may require some information to remain confidential. The regulatory challenge
is therefore to create meaningful accountability without unnecessarily exposing
sensitive operational information.
The study
also finds that grievance redressal deserves greater attention. Where an individual is incorrectly identified,
there should be a practical route to challenge the error and correct the
record. Accountability should not
begin only after serious harm has occurred.
13.
Findings
of the Study
The research
leads to the following findings.
First, FRT can serve
legitimate public purposes, but the constitutional assessment depends on how
and why it is used.
Second, privacy is a
central constitutional concern because facial recognition involves the
collection and processing of information connected to identifiable individuals.
Third, Article
14 becomes relevant
where automated systems
or administrative practices
produce arbitrary or discriminatory outcomes.
Fourth, Articles
19(1)(a) and 19(1)(b) may become relevant where widespread monitoring has a
chilling effect on speech or peaceful assembly.
Fifth, the
proportionality framework requires more than a general claim of public safety.
The State should be able to show lawful authority,
a legitimate objective,
a rational connection, necessity and a proportionate impact.
Sixth, existing
statutes address parts
of the broader biometric and data-governance environment, but they should not be described as a single, comprehensive FRT-specific regulatory framework.
Finally, legal
safeguards need to be accompanied by technical safeguards. Independent testing,
human review, access controls and mechanisms for correcting errors are essential
if automated identification is to be used
responsibly.
14.
Recommendations
1. Clear
statutory authority: Public deployment of FRT should have a clearly
identifiable legal basis and defined limits.
2. Purpose
limitation: Authorities should specify the precise purpose for which facial
data and recognition systems are used. Expansion to unrelated purposes
should require fresh legal justification.
3. Necessity
and proportionality assessment: Before deployment, authorities should document
why FRT is necessary and why less intrusive alternatives are inadequate.
4. Independent
oversight: Significant surveillance programmes should be subject to independent
review, auditing and periodic assessment.
5. Accuracy and bias testing:
Systems should undergo
independent testing using representative
datasets, with results and limitations documented for accountable decision-making.
6. Human
verification: A facial-recognition match should not, by itself, be treated as
conclusive proof of identity.
Human review should be required before serious adverse
action is taken.
7. Data
minimisation and retention limits: Collection and retention should be limited
to what is necessary for the specified lawful purpose, with secure deletion procedures.
8. Grievance redressal: Individuals should have an accessible mechanism to challenge incorrect identification and seek correction or other appropriate remedies.
9. Parliamentary
accountability: Large-scale surveillance programmes should be subject to
democratic scrutiny and clear institutional responsibility.
10. Periodic
review: Surveillance systems should not be treated as permanent simply because
they have been deployed.
Their necessity,
effectiveness and rights impact should be reviewed periodically.
15.
Discussion
The constitutional debate around AI-based
surveillance is ultimately a debate about the exercise
of public power.
The State may have genuine reasons
to use technology for public safety, but the presence
of a legitimate objective does not remove the need for legal
safeguards. The more intrusive the technology, the more carefully its purpose, scope and consequences should be examined.
FRT also illustrates
why constitutional analysis cannot be separated completely from technical
design. A legal rule requiring proportionality has little practical value if
the system is inaccurate, if there is no record of how a match was generated,
or if an individual has no way to challenge the result. Conversely, technical
accuracy alone does not make a surveillance system constitutionally
permissible. A highly accurate system can still be used for an unlawful or excessive purpose.
The appropriate approach is therefore
one of accountable innovation. Security
needs, technological
development and individual rights do not necessarily have to be treated as mutually exclusive. They can coexist
when surveillance is based on clear authority, narrowly
defined purposes, reliable
systems, independent oversight
and effective remedies.
16.
Conclusion
Facial Recognition Technology demonstrates both the possibilities and the difficulties created by AI-based surveillance. It can assist public
authorities in identification and investigation, but its ability to operate at
scale also creates serious questions concerning privacy, equality, liberty and
democratic participation.
In India, Articles 14, 19 and 21 provide
the principal constitutional framework for examining
these questions.
The
privacy jurisprudence beginning with Puttaswamy,
together with the broader principles developed in Maneka Gandhi, Anuradha
Bhasin and surveillance cases such as PUCL,
requires State action to be supported by law and justified through constitutional standards.
The research does not
suggest that every use of FRT should be treated in the same way. Instead, it
highlights the importance of purpose,
scale, necessity, safeguards and accountability. A targeted system
used for a
clearly defined and lawful objective
raises different questions
from continuous identification of everyone in a public space.
India would benefit from a clearer
rights-centred framework for AI surveillance that brings together statutory
authority, proportionality, data safeguards, technical auditing, human
review, independent oversight and effective remedies.
Such an approach would allow technological tools to support legitimate public
objectives while keeping their use within the boundaries of constitutional governance and the rule of law.
17.
References
1. Constitution of India.
2. Justice K.S. Puttaswamy (Retd.) v. Union of India, (2017) 10 SCC 1.
3. Maneka Gandhi v. Union of India, (1978) 1 SCC 248.
4. Anuradha Bhasin v. Union of India, (2020) 3 SCC 637.
5. People's Union for Civil Liberties v. Union of India, (1997) 1 SCC 301.
6. Kharak Singh v. State of Uttar Pradesh, AIR 1963 SC 1295.
7. Gobind v. State of Madhya Pradesh, (1975) 2 SCC 148.
8. Sadhan Haldar v. State (NCT of Delhi) &
Ors., W.P. (Crl.) 1560/2017, Delhi High Court orders (2019).
9. Criminal Procedure
(Identification) Act, 2022.
10. Digital Personal
Data Protection Act, 2023, especially
Section 17.
11. NITI
Aayog, Responsible AI for All: Adopting
the Framework – A Use Case Approach on Facial Recognition Technology.
12. Smriti
Parsheera, Adoption and Regulation of
Facial Recognition Technologies in India: Why and Why Not?, Data Governance
Network Working Paper 05.
13. Faizan
Mustafa & Utkarsh Leo, On Facial
Recognition and Fundamental Rights in India: A Law and Technology Perspective.
14. Rahul
Sharma & Virender Negi, Facial
Recognition Surveillance and the Indian Constitution: A Critical Analysis of
Privacy, Equality & State Power, International Journal of Law
Management & Humanities, Vol. 9, Issue 2 (2026).
Publication
note: This manuscript has been edited
for clarity, natural academic flow and consistency. It should still be
formatted according to the specific
journal's author guidelines before submission, including
its required citation
style, word limit, plagiarism declaration,
author-affiliation format and reference requirements.
How to Cite This Article
ROHAN V, KARTHIKEYAN S, BHARATH S, CONSTITUTIONAL CHALLENGES OF AI-BASED SURVEILLANCE IN INDIA., White Black Legal – International Law Journal, ISSN: 2581-8503, Vol. Volume 4, Issue VOLUME 4 ISSUE 2, October 2026, pp. 1209-1222. Available at: https://www.whiteblacklegal.co.in/details/constitutional-challenges-of-ai-based-surveillance-in-india
Author & Publication Record
Authors: ROHAN V, KARTHIKEYAN S & BHARATH S
Registration ID: 107198 | Published Paper ID: WBL7198, WBL7199 & WBL7200
Year: Oct- 2026 | Volume: 4 | Issue: 2
Approved ISSN: 2581-8503 | Country: Delhi, India
Page No.: 1209-1222
Full Text Preview
Open in New Tab
Copied