Constitutional Safeguards, Data Protection, Algorithmic Accountability and the Emerging Framework of AI Governance

By Rajiv Gupta, Advocate
Founder – KarVai Legal Solutions
AI Governance | Risk Management | EU AI Act | GDPR | FRIA

Professional Consultation: AI Governance, Fundamental Rights Impact Assessment (FRIA), EU AI Act Compliance, GDPR and Cross-Border Legal Advisory.

1. Introduction

Artificial Intelligence is fundamentally changing the manner in which governments, businesses and institutions make decisions affecting individuals.

AI systems are increasingly being deployed in recruitment, banking, insurance, healthcare, education, policing, public administration and judicial processes.

While these technologies offer considerable advantages in efficiency and decision-making, they also create serious concerns relating to privacy, equality, discrimination, freedom of expression, human dignity and access to justice.

The central legal question is no longer confined to whether an AI system functions efficiently or produces technically accurate results.

The more important question is:

Can an AI system make decisions affecting human beings without compromising their constitutionally and legally protected fundamental rights?

Both India and the European Union recognise the importance of protecting fundamental rights. However, their approaches to regulating the consequences of artificial intelligence demonstrate significant differences.

India principally relies upon constitutional guarantees, judicial interpretation, data protection legislation and sector-specific regulatory measures.

The European Union, on the other hand, has developed a more integrated regulatory structure combining the Charter of Fundamental Rights, the General Data Protection Regulation (GDPR) and the Artificial Intelligence Act (EU AI Act).

The introduction of Fundamental Rights Impact Assessment under Article 27 of the EU AI Act is particularly significant because it requires specified deployers of high-risk AI systems to examine the potential consequences of their deployment for individuals and groups.

This article examines the two legal systems, identifies their similarities and differences, and considers what India can learn from the European approach.

2. Constitutional Protection of Fundamental Rights in India

The Constitution of India provides a comprehensive framework for the protection of fundamental rights.

Part III of the Constitution, comprising Articles 12 to 35, establishes the fundamental rights available to individuals.

Several constitutional provisions have particular relevance to artificial intelligence.

A. Article 14 – Equality Before Law and Equal Protection of Laws

Article 14 guarantees equality before the law and equal protection of the laws.

This provision has considerable relevance to AI systems that make decisions concerning individuals.

Consider an AI-powered recruitment system used by a private or public employer.

The system evaluates applications based on educational qualifications, employment history, previous experience and other parameters.

Suppose the algorithm consistently gives lower scores to women who have taken career breaks or candidates belonging to particular social backgrounds.

Such outcomes may raise questions of discrimination and equality.

In the case of governmental decision-making, Article 14 provides a direct constitutional standard against arbitrary or discriminatory State action.

Its implications may also extend to private conduct through applicable legislation, constitutional doctrines and judicially recognised positive obligations.

The important question is whether an apparently neutral algorithm produces discriminatory outcomes.

An AI system need not expressly consider caste, religion, gender or other protected characteristics to create discriminatory consequences.

It may use indirect variables or statistical correlations that produce similar results.

This makes Article 14 particularly relevant to algorithmic fairness.

B. Article 15 – Prohibition of Discrimination

Article 15 prohibits discrimination by the State against citizens on specified grounds, including religion, race, caste, sex and place of birth.

It also permits affirmative measures for specified disadvantaged groups.

This constitutional provision becomes particularly relevant when AI systems are deployed in education, recruitment, public services and welfare distribution.

For example, an AI-based admission system may use historical academic data to determine eligibility for educational institutions.

If the underlying data reflects historical inequalities, the system may reproduce or reinforce existing disadvantages.

The challenge is to ensure that technological decision-making does not undermine constitutional commitments to substantive equality.

C. Article 19 – Protection of Fundamental Freedoms

Article 19 guarantees specified freedoms to Indian citizens, including:

  • Freedom of speech and expression.
  • Freedom to assemble peacefully.
  • Freedom to form associations.
  • Freedom of movement.
  • Freedom to reside and settle in any part of India.
  • Freedom to practise any profession or carry on any occupation, trade or business.

AI systems can affect these freedoms in several ways.

Automated content moderation may restrict lawful expression.

AI-based surveillance and behavioural profiling may discourage individuals from exercising their freedom of association or expression.

Automated systems used in professional licensing or employment may affect an individual’s ability to pursue an occupation.

Consequently, AI deployment by governmental authorities must be examined against the constitutional limitations applicable to restrictions on Article 19 freedoms.

D. Article 21 – Right to Life and Personal Liberty

Article 21 has perhaps the greatest significance in the context of AI governance.

The Supreme Court of India has interpreted the right to life and personal liberty broadly, recognising that it includes dimensions of human dignity, autonomy and privacy.

In Justice K.S. Puttaswamy (Retd.) v. Union of India (2017), the Supreme Court unanimously recognised privacy as a constitutionally protected fundamental right.

The judgment is particularly important in the age of artificial intelligence.

AI systems can collect, process, analyse and infer information about individuals on an unprecedented scale.

They may predict:

  • Financial circumstances.
  • Health conditions.
  • Personal preferences.
  • Behavioural patterns.
  • Social relationships.
  • Location and movement.
  • Individual vulnerabilities.

Even where an individual has not expressly disclosed certain information, an AI system may infer it from other available data.

This raises important questions concerning informational privacy, personal autonomy and individual dignity.

Article 21 therefore provides an important constitutional foundation for examining the human consequences of AI deployment in India.

3. Fundamental Rights Protection in the European Union

The European Union follows a different but equally important constitutional and legislative structure.

The principal instrument is the Charter of Fundamental Rights of the European Union.

The Charter consolidates various civil, political, economic and social rights into a common framework.

The following provisions are particularly relevant to AI governance.

Charter provisionProtected rightRelevance to AI
Article 1Human dignityPreventing degrading or exploitative AI applications
Article 7Respect for private and family lifeAI surveillance and behavioural profiling
Article 8Protection of personal dataCollection and processing of personal information
Article 11Freedom of expressionAI content moderation and automated censorship
Article 20Equality before the lawEqual treatment in automated decisions
Article 21Non-discriminationAlgorithmic discrimination
Article 23Equality between women and menGender bias in AI recruitment
Article 24Rights of the childAI systems affecting children
Article 26Integration of persons with disabilitiesAccessibility and inclusive AI
Article 47Right to an effective remedy and fair trialChallenging harmful AI-assisted decisions

An important distinction is that the Charter applies to EU institutions and Member States when they are implementing EU law, as provided under Article 51.

It does not automatically operate as a general constitutional code governing every private activity.

Nevertheless, EU legislation such as the GDPR and AI Act creates specific obligations for private organisations within their respective scopes.

4. India and the EU: A Comparative Constitutional Analysis

The following comparison illustrates the principal similarities and differences.

SubjectIndiaEuropean Union
EqualityArticle 14 of the ConstitutionArticles 20 and 21 of the Charter
Non-discriminationArticles 15 and 16, together with other applicable lawsArticle 21 of the Charter and equality legislation
PrivacyArticle 21, as interpreted by the Supreme CourtArticles 7 and 8 of the Charter
Human dignityRecognised through constitutional jurisprudence, particularly Article 21Expressly recognised under Article 1 of the Charter
Freedom of expressionArticle 19(1)(a)Article 11 of the Charter
Personal data protectionDigital Personal Data Protection Act, 2023GDPR
AI-specific regulationNo equivalent comprehensive horizontal AI ActEU AI Act
Fundamental Rights Impact AssessmentNo general statutory equivalent to Article 27Article 27 of the AI Act
Automated decision-makingNo equivalent general framework to GDPR Article 22GDPR Article 22
AI risk classificationSectoral and policy-based approachesRisk-based classification under the AI Act
Regulatory supervisionMultiple sectoral and statutory authoritiesAI Act market surveillance authorities and GDPR supervisory authorities

The comparison demonstrates that both jurisdictions recognise similar fundamental rights, but the manner in which these rights are translated into operational compliance requirements differs.

5. Data Protection: India’s DPDP Act and the European GDPR

Data protection is one of the most important areas of intersection between fundamental rights and artificial intelligence.

India’s Digital Personal Data Protection Act, 2023

India enacted the Digital Personal Data Protection Act, 2023, to regulate the processing of digital personal data.

The legislation recognises the importance of protecting individuals’ personal data while permitting lawful processing for legitimate purposes.

Its framework includes:

  • Obligations of Data Fiduciaries.
  • Consent requirements.
  • Notice to individuals.
  • Rights of Data Principals.
  • Duties of Data Fiduciaries.
  • Special obligations for Significant Data Fiduciaries.
  • Data Protection Board of India.
  • Provisions concerning cross-border transfers.

The Act is an important development in India’s data protection framework.

However, it is important to distinguish between data protection and comprehensive AI governance.

A data protection law does not, by itself, address every risk arising from algorithmic discrimination, automated decision-making, explainability or the deployment of high-risk AI systems.

The European GDPR

The GDPR provides a comprehensive framework for processing personal data within its scope.

Some of its particularly relevant provisions include:

  • Article 5: Principles relating to processing of personal data.
  • Article 6: Lawfulness of processing.
  • Article 9: Processing of special categories of personal data.
  • Article 13: Information to be provided where personal data are collected.
  • Article 15: Right of access.
  • Article 22: Automated individual decision-making, including profiling.
  • Article 25: Data protection by design and by default.
  • Article 32: Security of processing.
  • Article 35: Data Protection Impact Assessment.

Article 22 is particularly significant.

It establishes a right not to be subject to a decision based solely on automated processing, including profiling, where that decision produces legal effects or similarly significantly affects an individual, subject to specified exceptions and safeguards.

Where applicable, safeguards include the right to obtain human intervention, express one’s point of view and contest the decision.

This is important because it recognises that automated decisions can have serious consequences for individuals.

India’s DPDP Act does not contain an equivalent general provision to Article 22 GDPR.

This is an important area for future legislative and regulatory consideration.

6. The EU AI Act: Moving From Data Protection to AI Governance

The European Union has taken another significant step through Regulation (EU) 2024/1689, commonly known as the EU AI Act.

Unlike the GDPR, which principally regulates the processing of personal data, the AI Act establishes a risk-based framework governing the placing on the market, putting into service and use of AI systems within its scope.

Its regulatory structure includes:

  1. Prohibited AI practices.
  2. High-risk AI systems.
  3. Transparency obligations for certain AI systems.
  4. Obligations concerning general-purpose AI models.
  5. Governance and market surveillance mechanisms.

The Act recognises that certain AI applications can create unacceptable risks to fundamental rights, health, safety and democratic values.

Its high-risk framework is particularly important in areas such as:

  • Recruitment and employment.
  • Education and vocational training.
  • Access to essential services.
  • Creditworthiness assessment.
  • Life and health insurance risk assessment.
  • Law enforcement.
  • Migration and border management.
  • Administration of justice and democratic processes.

Article 27: Fundamental Rights Impact Assessment

Article 27 introduces a specific obligation upon certain deployers of high-risk AI systems.

Before the first deployment, covered organisations must assess the potential impact of the system on fundamental rights.

The assessment includes:

  • Description of the organisational processes.
  • Duration and frequency of use.
  • Categories of individuals affected.
  • Specific risks of harm.
  • Human oversight arrangements.
  • Mitigation measures and internal complaint mechanisms.

The significance of Article 27 is that it moves the responsibility for fundamental rights protection into the organisation’s actual deployment process.

It is no longer sufficient to examine whether an AI system has been technically developed in accordance with applicable requirements.

The deployer must also examine how the system affects people in its particular operational environment.

This is a major development in the practical implementation of human-centric AI governance.

It must, however, be remembered that Article 27 applies to specified categories of deployers and is not a universal FRIA obligation for every organisation using AI.

7. Why India’s Present Framework Requires Further Development in AI Governance

India has a substantial constitutional foundation for protecting fundamental rights.

However, the increasing use of AI creates new regulatory questions that require more specific treatment.

A. Absence of a General Statutory FRIA Requirement

Unlike Article 27 of the EU AI Act, Indian law does not currently establish a general statutory requirement for specified deployers of high-risk AI systems to conduct a dedicated Fundamental Rights Impact Assessment.

Organisations may undertake privacy assessments, internal risk assessments and other compliance exercises.

However, these are not equivalent to a general Article 27-style FRIA obligation.

B. Algorithmic Discrimination

India’s constitutional equality provisions provide a foundation for challenging discriminatory State action.

However, organisations also need practical procedures for identifying algorithmic bias.

For instance, an AI recruitment platform may unintentionally disadvantage candidates from certain educational, linguistic or socio-economic backgrounds.

There is a need for more specific guidance concerning algorithmic fairness testing, documentation and accountability.

C. Automated Decision-Making

The increasing use of automated decision-making creates important concerns.

For example:

  • Automated rejection of loan applications.
  • AI-based insurance risk assessment.
  • Automated recruitment screening.
  • AI-assisted allocation of welfare benefits.

India’s current DPDP framework does not establish a general equivalent of GDPR Article 22.

This leaves an important question concerning the procedural safeguards available to individuals affected by automated decisions.

D. Human Oversight

Human oversight is essential where AI systems influence decisions affecting people’s rights and opportunities.

An organisation should not be permitted to treat an automated recommendation as unquestionable merely because it has been generated by a sophisticated model.

Specific governance standards addressing the competence, authority and independence of human reviewers would be valuable.

E. Transparency and Explainability

Individuals should have meaningful information about consequential decisions affecting them.

An AI system that rejects an applicant for employment or financial assistance should not become an impenetrable decision-making mechanism.

The legal framework should encourage appropriate transparency, explanation and opportunities to challenge harmful outcomes.

F. Independent Technical and Legal Assessment

AI governance requires both legal and technical expertise.

An effective regulatory system must provide mechanisms for examining algorithms, datasets, system performance and the consequences of deployment.

This requires collaboration between legal professionals, computer scientists, data protection experts and regulators.

8. A Hypothetical Comparative Case Study: AI-Based Recruitment

Consider a multinational company operating in India and Germany.

It introduces an AI-powered recruitment platform to screen thousands of job applications.

The system evaluates:

  • Educational qualifications.
  • Employment history.
  • Professional experience.
  • Language proficiency.
  • Previous employment patterns.
  • Interview responses.

After several months, the company discovers that the system consistently gives lower scores to women who have taken career breaks.

It also appears to disadvantage applicants from certain educational institutions.

How would the two legal systems approach this situation?

Under the Indian Legal Framework

The company would need to consider:

  1. Whether any applicable constitutional obligations arise, particularly if State action is involved.
  2. Whether applicable employment and anti-discrimination laws have been violated.
  3. Whether personal data has been processed in accordance with the DPDP Act, once the relevant provisions become applicable.
  4. Whether the organisation has complied with other applicable contractual and sectoral obligations.
  5. Whether affected individuals have appropriate legal remedies.

Article 14 would be particularly relevant where governmental action is involved.

The constitutional principles of equality and non-arbitrariness would also inform the assessment of State action.

Under the European Legal Framework

The organisation would need to consider several additional regulatory requirements.

First: EU AI Act classification

The recruitment system would generally fall within the high-risk category under Annex III, point 4, subject to the applicable classification rules and exceptions.

Second: GDPR compliance

The organisation would need to examine the lawfulness of personal data processing, transparency, data minimisation and other GDPR obligations.

Article 22 would also require consideration if consequential decisions are based solely on automated processing and fall within its scope.

Third: Fundamental Rights Impact Assessment

If the organisation is a deployer covered by Article 27, it must assess the fundamental rights implications before first deployment.

Fourth: Non-discrimination

The organisation would need to examine applicable EU equality and employment laws, together with Articles 20 and 21 of the Charter where applicable.

Fifth: Human oversight

The organisation must ensure that human oversight arrangements are meaningful and effective under the applicable AI Act requirements.

Sixth: Corrective measures

The company should investigate the discriminatory outcomes, take appropriate mitigation measures and establish procedures for addressing affected individuals’ complaints.

This example demonstrates that the European framework provides a more specifically articulated AI compliance structure.

It also demonstrates why organisations operating in both jurisdictions need to understand the differences between constitutional protection and operational regulatory obligations.

9. Can India Adopt the European FRIA Model?

The European FRIA model offers several useful lessons for India.

However, India need not reproduce the EU AI Act in its entirety.

A calibrated approach could be considered.

1. Introduce High-Risk AI Impact Assessments

India could consider introducing a statutory requirement for organisations deploying specified high-impact AI systems to conduct impact assessments before deployment.

Such assessments could initially focus on:

  • Recruitment.
  • Healthcare.
  • Creditworthiness.
  • Insurance.
  • Education.
  • Public welfare.
  • Law enforcement.
  • Public administration.

2. Establish Algorithmic Accountability Standards

Organisations deploying consequential AI systems should maintain appropriate documentation concerning:

  • Data sources.
  • Intended purpose.
  • Model limitations.
  • Known biases.
  • Testing procedures.
  • Human oversight.
  • Corrective measures.

3. Strengthen Individual Rights

India could consider more explicit procedural safeguards for individuals affected by consequential automated decisions.

These may include:

  • Appropriate notice.
  • Meaningful explanations.
  • Human review.
  • Opportunities to contest decisions.
  • Effective complaint mechanisms.

4. Develop Human Oversight Guidelines

Organisations should be required to establish meaningful human supervision where AI systems influence important decisions affecting individuals.

5. Encourage Independent AI Audits

Independent legal and technical assessments could help identify discrimination, operational failures and governance weaknesses.

6. Integrate Privacy and Fundamental Rights Assessments

India could consider a framework integrating existing privacy assessments with broader fundamental rights assessments.

This would avoid unnecessary duplication while ensuring that AI-related risks beyond personal data protection are also addressed.

10. The Opportunity for Indian Legal Professionals in European AI Governance

The comparison between India and the European Union creates an interesting professional opportunity.

European organisations increasingly need professionals who understand the relationship between:

  • Constitutional law.
  • Fundamental rights.
  • Data protection.
  • Algorithmic accountability.
  • AI governance.
  • Regulatory compliance.
  • Human oversight.
  • Legal remedies.

A legal professional with substantial litigation experience can contribute meaningfully to this field.

FRIA is particularly interesting because it requires the ability to identify legal risks, analyse the potential consequences of organisational decisions and determine whether appropriate safeguards and remedies exist.

This is not exclusively a technical exercise.

It is fundamentally an exercise in legal reasoning, risk identification and institutional accountability.

Indian legal professionals familiar with constitutional jurisprudence, administrative law, equality principles, privacy and judicial remedies can develop relevant expertise in this emerging area.

However, they must also acquire a working understanding of AI systems, algorithmic risks, data governance and the specific obligations imposed by European legislation.

11. Conclusion: The Future of AI Must Be Human-Centric

India and the European Union share a fundamental commitment to protecting human dignity, equality, privacy and individual liberty.

Their legal systems, however, approach the challenges of artificial intelligence through different regulatory structures.

India’s Constitution provides a strong foundation for protecting fundamental rights, supported by an evolving data protection framework and judicial interpretation.

The European Union has gone further in establishing dedicated AI legislation that translates certain fundamental rights concerns into specific obligations for providers and deployers.

Article 27 of the EU AI Act is particularly significant because it recognises that fundamental rights protection must form part of the actual decision-making process of organisations deploying high-risk AI systems.

The future challenge for India is to bridge the gap between constitutional principles and practical AI governance.

For Europe, the challenge will be ensuring that the regulatory obligations already established translate into meaningful protection in real-world deployments.

Ultimately, the success of AI regulation should not be measured merely by the number of compliance documents prepared or technical standards satisfied.

It should be measured by whether individuals remain protected against arbitrary, discriminatory and unjust consequences of technological decision-making.

Artificial Intelligence must remain a tool serving humanity, rather than becoming an unaccountable authority over human beings.

At KarVai Legal Solutions, we believe that the future of AI governance must be built upon the integration of constitutional values, fundamental rights, legal accountability and practical regulatory compliance.

Rajiv Gupta, Advocate
Founder – KarVai Legal Solutions
AI Governance | Risk Management | EU AI Act | GDPR | FRIA

Professional Consultation: AI Governance, Fundamental Rights Impact Assessment (FRIA), EU AI Act Compliance, GDPR and Cross-Border Legal Advisory.