.jpg)
Privacy, Policing, and the Algorithm: Rethinking Surveillance and Data Protection in the Age of AI
Updated: Sep 27
Paper Code: AIJACLAV30RP2025
Category: Research Paper
Date of Publication: Feb 10, 2026
Citation: Jashan Sharma, “Privacy, Policing, and the Algorithm: Rethinking Surveillance and Data Protection in the Age of AI", 5, AIJACLA, 332, 332–341 (2025), <https://www.aequivic.in/post/privacy-policing-and-the-algorithm-rethinking-surveillance-and-data-protection-in-the-age-of-ai>
Author Details: Mrs. Jashan Sharma, Student, LL.M. (Criminal Law) 1st year, Amity Law School, Amity University, Mohali, Punjab
Abstract
The architecture of state authority and surveillance has changed significantly due to the growing use of artificial intelligence in modern policing and governance. Technologies such as large-scale data fusion centers, real-time facial recognition systems, and predictive policing algorithms allow law enforcement agencies to monitor public spaces, predict criminal behaviour, and identify individuals with unprecedented precision and scale. Even if such technologies aim at improving the efficiency of the work and enhancing security, they put in jeopardy privacy rights, rights of due process, and democratic accountability in a very serious manner. The constitutional recognition of privacy rights through Justice K.S. Puttaswamy's landmark decision contrasts very much with the fast spreading of algorithmic monitoring systems that are hardly checked in India.[1] While global norms such as the General Data Protection Regulation (herein after GDPR) provide robust safeguards, a recent study reveals their limitations in addressing bias and lack of transparency in machine learning algorithms.
The Digital Personal Data Protection Act of India (herein after DPDP) 2023 has allowed India many shortcomings with regards to the biometric technologies and the algorithmic policing tools due to the fact that it gives a wide range of exemptions to state agencies. In this way, the present study deals with the democratic challenges posed by algorithmic law enforcement, illuminates the shortcomings of the DPDP Act 2023, assesses its consistency with international standards, and proposes the necessary models and doctrines for the effective governance of the artificial intelligence era by investigating the ethical, legal, and constitutional issues of AI-enabledsurveillance.[2]
Keywords:- AI surveillance, privacy rights, biometric technologies, predictive policing algorithms, facial recognition systems, DPDP Act, 2023.
Paper Code: AIJACLAV30RP2025
1. INTRODUCTION
The constitutional basis for privacy in India can be traced down to the three Articles 14, 19, and 21 of the Indian Constitution.The 2017 Puttaswamy verdict changed the configuration of the Indian constitution profoundly by explicitly identifying privacy as a core right necessary for human dignity and individual freedom. The doctrinal foundation from this case for the comprehension of such problems is the proportionality principle.[3]
But, quantum surveillance threatens the existence of these constitutional safeguards.[4]Quantum surveillance is not only about the amount of data that is collected but it is also about a complete change in the nature of state power due to AI (herein after Artificial Intelligence) systems.[5]
The DPDP Act has certain limitations. The Act, in its Section 17, provides for a large number of exceptions to state agencies and other government bodies when they are dealing with personal data for the purposes of national security, sovereignty, maintenance of public order, prevention of crime, etc.[6]
Encryption technologies are tools that users can utilize to keep their online communication and personal data safe from the government's spying activities.[7] However, in a post-encryption world, it is a completely new set of problems that law enforcement and national security agencies face.[8]
2. AI OVERSIGHT AND THE DEATH OF ANONYMITY
The development of AI surveillance technologies combined with the rapid deployment of sensor networks severely undermines the possibility of remaining anonymous in public spaces.[9] They capture people's faces in images, extract the unique facial features, and then look up the features in the databases for matches to identify the person with the highest degree of confidence.[10]
Gait recognition is an emerging issue that is less talked about than face recognition but might eventually turn out to be more intrusive than facial recognition.[11]. Gait recognition can pinpoint the identity of the person in a case where rapid movement, great distance, and face-covering is involved, and also can function in dim light, fog, or similar conditions where facial recognition is not.[12]
Voice biometric systems identification of individuals based on the unique aspects of the speech, such as pitch, resonance, rhythm, and accent. While working with AI systems that can understand speech in different languages and dialects, voice biometric identification can identify any individual even in a linguistically diverse group. These systems are capable of recognizing the voice from a recorded telephone conversation, video recording, or even background audio at a surveilled location.
These biometric identification technologies have the capability to change the whole scenario of individuals' anonymity from them being unidentified in the public to becoming tracked and constantly monitored identifiable subjects. The right to anonymity in public is one of the basic democratic safeguards. Democracy calls for the need of being able to participate in political discussion, form associations, and do activities without the fear of authorities reporting on them.[13] Whistleblowers, investigative journalists, and human rights activists, to mention a few, are dependent on the ability of being anonymous or the ability to control who knows about their actions and connections.
Voice recognition technology is a source of threat especially regarding phone calls and private talks. The authorities can use voice biometric identification to break down the recorded conversations for identifying the speakers based on their voice traits. The technology may be used for identifying individuals who are reaching out to the civil rights organizations, religious groups, and political opposition parties or any such group that the state might be retaining under surveillance.[14]
The analysis of fingerprints can perform the analysis of prints taken from the scenes of the crime and can further find matches from the large-scale databases. The police have been using fingerprint identification as a tool for solving crime for many years, however, the creation of fingerprint databases that include not only arrested people but also unconvicted detainees for immigration, innocent witnesses, and in some cases, people who have not been involved, has given rise to new problems.[15]
Articles 21 and 29 of the Constitution: is for Dignity and the Right to Obscurity. The right to anonymity is based on the several provisions of the constitution. In the Puttaswamy case, it was decided that the right to privacy is an essential element of the right to life and personal liberty. Living a private life, which is free from state surveillance and intrusion, is essential to human dignity. Being anonymous in public is one of the components of the bigger right to privacy.[16]
Article 29 of the Constitution is a protector of minority rights and lists the rights of minority religious and linguistic groups. The right to religion includes going to services, gathering for worship, and participating in religious activities. Nevertheless, in terms of surveillance systems that intend to monitor individuals' participation in religious activities, they are a risk to this constitutional right. [17]
The idea of a right to obscurity has largely been derived from European law, mainly from the legal traditions of France and the EU. The right to obscurity is a concept whereby people are allowed to keep certain parts of their lives out of the public eye as well as from the state's scrutiny. The right to obscurity is a means for people to become and develop and still not be identified by their past deeds or associations. [18]
Both the EU and France have been very explicit in their acknowledgment of the right to obscurity through their data protection laws. The GDPR is one of the entities that recognizes the concept of the right to obscurity by including the right to be forgotten as one of its provisions. The French law has gone beyond the idea of the right to obscurity by accepting it as a necessary factor for personal development and freedom. [19]
3. LEGAL FRAMEWORK, INTERNTIONAL STANDARDS
AND CONSTITUTIONAL FONDATIONS
3.1The Justice Srikrishna committee and Data protection framework development
The Justice Srikrishna Committee was the need to answer India's data protection and privacy challenges. A committee, in the wake of the Puttaswamy verdict, was formed to address the exigency of legislating to protect the digital personal data of individuals. Prior to this, data protection in India was mainly regulated through the IT Act of 2000 with the addition of the Sensitive Personal Data or Information Rules of 2011.[20]
The framework failed to address issues such as consent, data minimization, purpose limitation, or the rights of data subjects to access, correct, or delete their personal data. The Srikrishna Committee on Justice consisted of legal experts, technology professionals, members of the civil society, and government officials, who after detailed discussions, came up with a design for a thorough data protection system.
The committee's philosophy was that the law should strike a balance between individuals' privacy rights and the need of organizations to use data for legitimate business and those of the state for national security and public order. After intense discussions, the committee came up with a thorough white paper followed by a draft Data Protection Bill reflecting its understanding.[21]
This Committee, referred to as the Committee of Experts on Data Protection Issues, recommended the comprehensive approach as the characteristic feature of privacy protection. The committee equally focused on the point of individual rights and empowerment. Its designed structure contained the rights of the people to have access to their personal data, to request that errors be corrected, to remove junior data, and to seek remedies when their rights have been violated. The assumption was that in order for privacy to be effectively safeguarded, data subjects had to be given control over their data instead of being passive objects of data collection.
3.2 The Supreme Court Puttaswamy decision: detailed analysis
The 2017 Justice K.S. Puttaswamy (Retd.) v. Union of India case is a landmark Indian constitutional ruling of its time. The case was a result of allegations to the Aadhaar program, a biometric identification system in India that had completed the registration of more than a billion citizens prior to the enforcement of any comprehensive data protection law. Retired Justice K.S. Puttaswamy filed a writ petition asserting that the Aadhaar scheme was in violation of the fundamental rights of the citizens under the Indian Constitution.
The SC constituted a nine-judge bench to address the question of whether the Constitution of India confers a fundamental right to privacy and the extent and exceptions of that right. The bench consisted of Chief Justice J.S. Khehar and Justices Jasti Chelameswar, S.A. Bobde, D.Y. Chandrachud, and others. The issue necessitated a Supreme Court revision of the decisions of the SC in earlier cases M.P. Sharma v. Satish Chandra (1954) and Kharak Singh v. State of Uttar Pradesh (1962), which had held that privacy is not a constitutional right. According to the majority opinion, Justice D.Y. Chandrachud argued that privacy is not separate from other fundamental rights listed in Part III of the Constitution but is the central idea of human dignity and personal freedom. Chandrachud highlighted the informational aspect of privacy by stating that people should have the right to control their personal data, choose the person who can get access, and the way which should be used. The opinion recognized that privacy is not only protection against physical intrusion but also the safeguarding of informational self-determination. The decision established a test with three prongs for the lawfulness of state action in respect of the right to privacy. The first is that the action must be 'by law' which means, among other things, that it has to be authorized by a valid statute or constitutional provision. The second one indicates that it should serve a legitimate state purpose, e.g., national security, prevention of crime, or protection of public health. Third, if the act passes the proportionality test, it means that it should be essential to achieving the legitimate aim and that the degree of invasion should not exceed what is necessary.
The Puttaswamy decision has overruled the previous rulings opposing the recognition of privacy as a fundamental right and has explicitly affirmed that privacy enjoys protection under Articles 14, 19, and 21 of the Constitution. It stressed that privacy is not only the right of a few but a prerequisite of human dignity. Protecting privacy, the court has recognized, is particularly imperative for the most vulnerable and marginalized groups, who, because of their condition, are at the co-existence of discrimination and abuse.[22]
The verdict echoed the point that the issue of privacy arises from the very act of collecting personal data, more especially biometric data, without any further stepsIt accepted that the availability of extensive databases with personal information, in conjunction with other data sources, can result in new ways of surveillance and control.[23]
3.3 OECD Ai principles and international implementation.
The OECD (Herein after Organisation for Economic Co-operation and Development) has outlined a set of principled guidance for AI that could be seen as a global regulatory framework reflecting the values and principles of advanced democracies. These principles were presented in 2019, and they represent the common vision of leading democratic countries regarding the governance of AI systems. Knowing that AI can raise privacy, fairness, transparency, and accountability issues, the AI principles argue that these problems should be addressed in the governance framework.[24]
Various countries like, Singapore has designed reliable measures of AI supervision in harmony with the OECD norms. Besides that, Canada, Australia, and a few other developed democratic countries, have set up management systems based on the OECD principles. Nevertheless, India is yet to adopt these principles in the formulation of its AI governance.[25]
The connection between the OECD AI principles and present UNESCO ethics and AI projects is not very clear.[26] In fact, UNESCO (herein after United Nations Educational, Scientific and Cultural Organization) has progressively been concentrating on formulating ethical standards for AI while realizing that worldwide cultural values and viewpoints have to be the main guiding factors in AI governance. Consequently, UNESCO emphasizes the need for such technologies to be utilized in the developing countries in a manner which is helpful to their culture instead of just sending the governance framework of the developed countries.[27]
3.4 UNESCO ethics of Ai report 2021: Comprehensive analysis
The 2021 UNESCO Report on the Ethics of Artificial Intelligence is a global collaborative effort to create ethical frameworks for the governance of AI that are not only based on the morals of one culture or region but, rather, consider the views of different cultures and regions. The report resulted from numerous consultations involving representatives from more than one hundred countries, NGOs, academic institutions, and the private sector.[28]
According to the UNESCO report, four core values should be the basis of global governance of AI. The first, it states that AI systems should not be the cause of human autonomy losing their power, but rather they should be the ones to complement human autonomy. In no way should it be that AI systems dictate to humans, but on the contrary, human beings should be the ones to decide by means of AI systems. The second, which means that AI systems must be created in such a way that they benefit human well-being and keep individuals safe from any kind of harm. And this covers mental health, physical safety, social relationships, and freedom. The third set of values are equity and non-discrimination to be sure of AI systems not embedding or exacerbating existing inequalities. AI systems should be evaluated for discriminatory effects and if any, they should recode their operations until the effect is neutralized. The fourth value solidarity and inclusivity are the key to have AI governance respectful of the community and taking into account the people's opinions and by doing so, they will be sharing the benefits and the drawbacks in a fair way.[29]
The UNESCO report enumerates the principles of AI governance as diverse strategies: [30]
I. Firstly, transparency and explainability are of utmost importance. The AI systems should be structured in a way that the people impacted can grasp their functioning. Moreover, the choices reached by these systems ought to be not only evident but also defensible.
II. Secondly, the role of humans and their supervision along with the intervention is very essential here. By this it is meant that control exercised in a significant way by the human over particularly those decisions that affect basic rights is a must condition.
III. Thirdly, the notion of sustainability and environmental friendliness cannot be overlooked. It is only fair that the AI machines and innovations should neither be a source of big environmental pollutions nor should they consume huge amounts of non-renewable resources.
IV. Fourthly, privacy needs to be safeguarded at all times. Setting out definite rules on each step of personal data handling operation, including collection, use, and security, is imperative especially in the case of AI systems.
V. Fifthly, accountability and responsibility constitute an important part of the whole ensemble. There should be unambiguous directions on who and to what extent is responsible for the effects of AI and the ways to tackle such problems, if they arise, should also be put forward.
VI. Sixthly, an utmost necessity is the detection and alleviation of bias. During the creation of AI systems they must be checked to ensure that they are bias-free and even when these are put to use in the field, there should be unceasing monitoring of any biased results’ occurrence.
VII. Seventhly, accessibility and digital inclusion are of great importance concepts. Herein lies the point that no AI system should be to the disadvantage of differently abled individuals and at the same time the existence of some digital divides should not be a reason for any person to be deprived of the accessing of AI services.
VIII. Eighthly, one should not underestimate the value of security and robustness. AI systems need to be safeguarded against the intentions of people who want to do them harm and, on the other hand, these systems should be able to counter the unexpected input and situations and still deliver their services reliably.
IX. Ninthly, environmental responsibility should be a top priority for all concerned. That is the only way the society can benefit from AI systems without causing serious irreversible damage to the environment and also guarantee the sustainability of the development.
X. Tenthly, ethical oversight and governance should be looked upon as a necessity rather than an option. An ethical review of AI systems should be a routine exercise and the management frameworks have to be there to ensure that the compliance of these systems.[31]
3.5 Lack of data fiduciary duties Comparable to OECD guidelines
3.5.1 The loopholes in the DPDP ACT are:
· It lays down a definite provision for purpose limitation, according to which personal data can be processed only for lawful and specific purposes that have been informed to the Data Principal.
· It is a storage limitation instrument that enforces regulations, wherein data fiduciaries are allowed to retain personal data only for the time that is necessary to achieve the purpose stated.
· The law prescribes thorough data minimization obligations and allows only such personal data to be collected that is indispensable for the processing purpose.
· It is equipped with provisions that forbid secondary data usage, thus, no processing beyond the original purpose is allowed if there is no new consent or legal ground.
3.5.2 OECD Privacy Principles Demand:
· The principle of limited collection puts a limit on the gathering of personal data to it being only through legal and fair methods and restricted to what is necessary.
· The data quality principle is that personal data have to be accurate, complete, and kept up to date.
· The specific purpose principle is that data collection should be only for clearly defined and legitimate purposes.
· The limited use principle limits the use and disclosure of personal data to the purposes that were indicated at the time of collection only.
· The security measures principle is that organizations have to protect personal data through reasonable safeguards against loss, unauthorized access, or misuse.
· The openness principle is that data handlers have to be transparent about their privacy policies and practices.
· The individual participation principle is the right of individuals to access, verify, and request correction of their personal data.
· The accountability principle is that the data controllers are responsible for ensuring that the privacy principles are observed.
What Required To Be Added:
· High-risk data processing should have been accompanied by mandatory Data Protection Impact Assessments (DPIAs).
· The purpose limitation has to be very specifically defined without any open-ended "deemed consent" clauses.
· There should be very clear restrictions on the secondary use of data that are in line with the OECD standards.
3.5.3 Broad “Deemed Consent” Provisions Loophole:
“Deemed consent” is a mechanism that allows data to be processed without an explicitly given consent in several instances such as in employment, public interest, government benefit delivery.
Essentially, this feature undermines the principles of purpose limitation and consent autonomy.
3.5.4 No Protection for Non-personal, Anonymized, or Sensitive Metadata Loophole:
· The Act is limited to "digital personal data" only.
· Metadata, anonymized, and inferred data, which are essentially AI profiling's fuel, have been left without proper protection.
Major norms in this respect of global standards :
· OECD and EU treat inferred and pseudonymized data as very sensitive ones.
What should have been added:
. • The rules for inferred data that are clear should have been added in order to regulate how insights drawn from personal information may be generated and used.
• It should have been included the clear rules for behavioral data to govern the collection and analysis of user actions and digital footprints.
• Clear rules for algorithmic predictions should have been given to facilitate transparency and accountability in automated decision-making.
• Clear rules for data gathered from public CCTV or IoT should have been laid down to ensure that the large-scale, ambient surveillance data is not misused.
3.5.5 No Specific Protections for Children's Data Beyond Consent Loophole:
· Targeted advertising to minors is not limited in any way, thus children are exposed to manipulative digital marketing without sufficient protection.
· There are no limitations on the use of surveillance tools for children, which means that technologies for monitoring children are advancing without being checked.
· There are no protections against AI profiling of students through EdTech platforms, which means that minors are exposed to automated judgments without being safeguarded.
What should have been added:
· If there had been a prohibition on AI profiling of children, it would have been intended to prevent the making of automated inferences about the abilities, behaviour, or personality of minors.
· Imposing a ban on behavioral advertising aimed at children should have been the requirement in order to stop the manipulative targeting that is based on the online activities of children.
3.5.6 Weak Data Localization and Cross-Border Transfer Protection Loophole:
· The authorities are free to approve the countries for data transfer by themselves without conducting thorough privacy checks.
· There are no standard contractual clauses or adequacy tests similar to those in GDPR.
What Should Have Been Added Clear adequacy standards that include:
• the principle of rule of law,
• the strong privacy protections,
• the independent regulators.
4. PROPOSED FRAMEWORKS FOR ETHICAL AI OVERSEE IN INDIA
In order for India to have a successful AI governance framework, practical legislative and technical requirements must be combined with constitutional protections. The rights to equality, privacy, and non-discrimination are established by India's constitutional framework as fundamental values rather than alternative policy options.
4.1 Integrating Constitutional Values
This examination makes sure that the systems satisfy the proportionality standard set forth in Justice K. S. Puttaswamy's seminal SC decision (2017). This test asks:
Does the AI system have legal authorisation?
Does it serve a valid state purpose, like administrative effectiveness or public safety?
Is the degree of violation of individual rights appropriate and required to accomplish the justifiable goal?
The deployment of AI systems should be prohibited if they fail any portion of this test. According to this idea, the first hurdle of AI control is compliance with the constitution which recognises court precedence.
4.2 Mandatory Bias Auditing and Testing
Bias audits are a crucial part of the governance system that aims to stop and address the discriminatory effects of AI, especially against those groups protected by the constitution. The framework requires:
Extensive bias testing before the system is put to work, with biassed-patterned systems being adjusted or stopped;
Continual monitoring after deployment to locate newly arisen biases caused by changing data or populations;
Corrective measures, like re-calibrating the system, interrupting use or revoking licenses in case of persistent prejudice.
With this method of bias management, equality and fairness are directly incorporated into organisational and technical AI governance.
5. CONCLUSION AND FUTURE DIRECTIONS
With emerging AI and quantum surveillance technologies that could either dominate or be controlled by constitutional safeguards, India finds itself at a very decisive moment. It has been a great move by the Puttaswamy verdict to prepare the ground by confirming privacy as a right of the people and ensuring protection against the government's arbitrary surveillance.
On the other hand, acts like the DPDP that have an idea or two that contradict the established basis by, for example, exempting state agencies from accountability and not being comprehensive enough in regulating technologies such as facial recognition and predictive policing negate the effect of this security measure to a large extent.
5.1 Advancing Law Enforcement and Safeguarding
The DPDP legislation should be reformed in such a way that it assures compliance of any state surveillance with the Puttaswamy proportionality test.[32] Clear statutory norms for algorithmic transparency, regular bias audits, and independent supervision; Elimination or very significant reduction of the broad exemptions that currently facilitate government bodies in evading data protection rules.
5.2 The formation of an independent regulatory agency
In order to achieve dependable AI governance, there is a necessity for an unbiased regulatory body with a profound understanding of technology. This body should be empowered to:
• Conduct extensive reviews of algorithmic methods in order to ensure their compliance with statutory and constitutional provisions;
• Cease the implementation of AI applications that result in fundamental rights violations;
• Impose heavy punishments and sanctions on offenders;
• Oversee grievance mechanisms and be responsible for transparency in public communication.
The enhancement of public trust and accountability will be the result of such a body being able to act decisively if it is given the power to do so.
5.3 Recognizing the Right to Obscurity and Encryption
In order to protect encryption as an aspect of the right to privacy, the imposition of government orders for backdoors or mandatory key disclosures that would result in compromised digital security should be disallowed. Strong encryption is a tool that can fend off the undesirable surveillance and is thus a vehicle for individual freedom.
5.4 The Efficiency of Resisting the Supremacy of Technological over Rights
Without this equilibrium, technology may result in continuous surveillance and discriminatory practices.
The Puttaswamy verdict lays down the principles that are crucial: the use of technology has to be in line with respect for constitutional rights and not the opposite.[33] The main objective is to make sure that technological systems do not undermine constitutional rights by going beyond legal oversight. The presence of comprehensive judicial and legislative supervision, an impartial regulatory authority and an active civil society is vital in order to preserve constitutional standards.
[1] K.S. Puttaswamy (Retd.) v. Union of India, (2017) 10 S.C.C. 1 (India)
[2] FAIZAN MUSTAFA, THE CONSTITUTION AND RELIGIOUS FREEDOM IN INDIA 234-267 (2015); Arvind P. Subramanian, Algorithmic Governance and the Fundamental Rights, 15 NUC. CONST. REV. 45-78 (2020)
[3] supra note 1 at 7
[4] Shoshana Zuboff, The Age of Surveillance Capitalism (2019)
[5]Patricia Aufderheide, Book Review (reviewing JULIE E. COHEN, BETWEEN TRUTH AND POWER: THE LEGAL CONSTRUCTIONS OF INFORMATIONAL CAPITALISM) 13 Int’l J. Comm. 5898 (2019)
[6] THE DIGITAL PERSONAL DATA PROTECTION ACT, 2023 (NO. 22 OF 2023)
[7] Bruce Schneier, Applied Cryptography: Protocols, Algorithms, and Source Code in C 3–12 (John Wiley & Sons 1996)
[8] Julie E. Cohen, Between Truth and Power: The Legal Constructions of Informational Capitalism 215–230 (Oxford Univ. Press 2019)
[9] supra note 35 at 18
[10] National Institute of Standards and Technology (NIST), Face Recognition Vendor Test (FRVT) Part 3: Demographics, 2–5 (2020)
[11] RecFaces, What Is Gait Recognition (on file with RecFaces)
[12] FENG HU, XIAOMING DU & GEYONG MIN, ADVANCED BIOMETRIC TECHNOLOGIES FOR SURVEILLANCE AND IDENTIFICATION 234-267 (2d ed. 2018)
[13] Innovatrics, Biometric Identification Glossary, https://www.innovatrics.com/glossary/biometric-identification/ (last visited on 05.10.2025)
[14] ELECTRONIC FRONTIER FOUNDATION, THE VOICE OF THE FUTURE: THE PRIVACY IMPLICATIONS OF VOICE RECOGNITION TECHNOLOGY 123-156 (2020)
[15] Fingerprints Analysis in Forensic Science, LIFS (last visited Dec. 5, 2025), https://lifs.co.in/blog/fingerprints-analysis-in-forensic-science.html (last visited on 11.10.2025)
[16] supra note 1 at 7
[17] Constitution of India art. 29.
[18] Guy Canivet, The French Constitution and European Union law: An Approach through the complexity of legal power relations, Revue Européenne du Droit, Issue #3 (Dec. 2021), available at https://geopolitique.eu/en/articles/french-constitution-and-european-union-law-approach-through-the-complexity-of-legal-power-relations/ (last visited 19.10.2025)
[19] LUCIANO FLORIDI & JOSH COWLEY, THE ETHICS OF ARTIFICIAL INTELLIGENCE: PRINCIPLES, CHALLENGES, AND OPPORTUNITIES 189-234 (2014); see also FRENCH DATA PROTECTION AUTHORITY, THE RIGHT TO OBSCURITY IN THE DIGITAL AGE: LEGAL ANALYSIS AND POLICY RECOMMENDATIONS 45-78 (2017)
[20] JUSTICE B.N. SRIKRISHNA, COMMITTEE OF EXPERTS ON DATA PROTECTION ISSUES, WHITE PAPER ON DATA PROTECTION IN INDIA (2017)
[21] Committee of Experts on a Data Protection Framework for India, Committee Report on Draft Personal Data Protection Bill, 2018 (“Justice Srikrishna Committee Report”) (2018),https://prsindia.org/files/bills_acts/bills_parliament/2019/Committee%20Report%20on%20Draft%20Personal%20Data%20Protection%20Bill,%202018_0.pdf (last visited on 11.10.2025)
[22] supra note 1 at 7
[23] supra note 1 at 7
[24] OECD, Artificial Intelligence in Society 15 (2019), https://doi.org/10.1787/eedfee77-en (last visited on 11.10.2025)
[25] ORGANIZATION FOR ECONOMIC COOPERATION AND DEVELOPMENT, OECD AI PRINCIPLES: RECOMMENDATIONS ON ARTIFICIAL INTELLIGENCE (2019) https://www.oecd.org/content/dam/oecd/en/publications/reports/2019/06/artificial-intelligence-in-society_c0054fa1/eedfee77-en.pdf (last visited on 20.10.205)
[26] Ibid at 28
[27] UNESCO, Recommendation on the Ethics of Artificial Intelligence https://www.unesco.org/en/artificial-intelligence/recommendation-ethics (last visited on 20.10.2025)
[28] Evelyne Para, The UNESCO Recommendation on the Ethics of Artificial Intelligence (July 18, 2024) https://www.soroptimistinternational.org/2024/07/18/the-unesco-recommendation-on-the-ethics-of-artificial-intelligence/ (last visited on 20.10.2025)
[29] UNESCO, Input in Reply to the OHCHR Report on Human Rights Council Resolution 47/23: Recommendation on the Ethics of Artificial Intelligence (Mar. 2022)
[30] UNITED NATIONS EDUCATIONAL, SCIENTIFIC AND CULTURAL ORGANIZATION, REPORT OF THE STUDY ON ETHICS OF ARTIFICIAL INTELLIGENCE 156-189 (2021)
[31] supra note 61 at 29
[32] supra note 1 at 7
[33] supra note 1 at 7