Non-human cognition doctrine accountability

1. Introduction

Non-Human Cognition Doctrine Accountability is an emerging concept in artificial intelligence law, constitutional governance, technology regulation, and legal responsibility. It concerns the establishment of accountability mechanisms for decisions, actions, and outputs generated by non-human cognitive systems, including artificial intelligence, autonomous machines, advanced robotics, neural networks, and other computational technologies.

Modern technological systems increasingly perform activities traditionally associated with human cognitive abilities, such as reasoning, learning, prediction, decision-making, language processing, and problem-solving. These systems are being introduced into healthcare, financial services, public administration, education, transportation, employment, and judicial support systems.

However, when a non-human cognitive system produces a harmful, discriminatory, inaccurate, or unlawful outcome, an important legal question arises: who should be held accountable?

For example, if an AI-based recruitment system rejects qualified candidates because of biased training data, responsibility may involve the organization deploying the system, its developers, data providers, or other parties involved in its operation.

The Non-Human Cognition Doctrine Accountability framework seeks to ensure that technological autonomy does not create a legal accountability gap. Its central principle is that decisions generated by artificial or non-human cognitive systems must remain subject to human oversight, legal scrutiny, transparency, and appropriate remedies.

It is important to clarify that this is an emerging analytical and normative concept rather than an independently recognized legal doctrine under Indian law.

2. Meaning and scope of non-human cognition accountability

Non-human cognition refers to computational or artificial systems capable of performing functions associated with human cognitive processes. These may include machine learning, automated reasoning, pattern recognition, predictive analysis, and autonomous decision-making.

Accountability refers to the legal and institutional obligation to explain, justify, review, and accept responsibility for the consequences of such systems.

The doctrine can be understood through the following dimensions.

A. Accountability for autonomous decisions

AI systems may generate decisions without direct human intervention at every stage. However, the absence of immediate human involvement should not eliminate the possibility of legal responsibility.

Organizations using these systems must establish mechanisms for reviewing consequential decisions.

B. Human responsibility

Artificial intelligence should not become a convenient means for organizations to avoid legal obligations.

Developers, manufacturers, operators, and deploying institutions may have responsibilities depending on their respective roles, control, conduct, and applicable legal rules.

C. Algorithmic transparency

Organizations should maintain appropriate information about how consequential AI systems operate, what data they use, and how their outputs influence decisions.

D. Protection against discriminatory outcomes

Non-human cognitive systems must not reproduce unlawful discrimination through biased datasets, defective algorithms, or inappropriate decision-making criteria.

E. Accountability for technological harm

Where AI systems cause financial, physical, psychological, or other legally recognized harm, applicable liability frameworks should provide appropriate mechanisms for investigation, compensation, and corrective action.

3. Constitutional foundation of non-human cognition accountability

The Indian constitutional framework provides important foundations for regulating AI-generated decisions.

A. Article 14 – Equality before law

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

Government institutions using AI systems must ensure that automated decisions do not produce arbitrary or discriminatory outcomes.

For example, an AI-based public-benefit allocation system should not unfairly exclude eligible individuals because of biased algorithmic classifications.

B. Article 19 – Fundamental freedoms

Article 19 protects several fundamental freedoms, including freedom of speech and expression and the right to practise a profession.

AI-based content moderation, automated employment restrictions, and algorithmic decisions affecting professional opportunities may raise constitutional concerns where they unjustifiably interfere with protected freedoms.

C. Article 21 – Life and personal liberty

Article 21 protects life, dignity, privacy, and personal liberty.

AI systems processing sensitive personal information or making consequential decisions affecting individuals must operate consistently with applicable privacy and constitutional safeguards.

D. Article 32 – Constitutional remedies

Article 32 provides the right to approach the Supreme Court for enforcement of fundamental rights.

Where governmental use of AI violates fundamental rights, affected individuals may seek appropriate constitutional remedies.

E. Article 226 – Judicial review

High Courts possess jurisdiction to issue appropriate writs for enforcement of fundamental rights and for other legal purposes.

This jurisdiction can provide an important mechanism for reviewing unlawful administrative decisions involving automated systems.

4. Important judicial decisions and case laws

There is no established Indian Supreme Court judgment expressly recognizing a Non-Human Cognition Doctrine Accountability. Nevertheless, the following judgments provide significant legal principles for developing AI accountability.

1. State of Punjab v. Gurdev Singh (1991)

The Supreme Court considered the legal consequences of an allegedly void administrative order and the importance of challenging unlawful decisions through appropriate legal proceedings.

Relevance: Government decisions generated or supported by automated systems must remain subject to legal review. Technological involvement cannot automatically validate an otherwise unlawful administrative decision.

2. Justice K.S. Puttaswamy v. Union of India (2017)

The Supreme Court recognized privacy as a fundamental right under Article 21.

The judgment emphasized informational privacy, individual autonomy, dignity, and protection against unjustified interference.

Relevance: AI systems frequently collect and process large quantities of personal information. Organizations must respect privacy rights and establish appropriate safeguards when using personal data for automated decision-making.

3. Shreya Singhal v. Union of India (2015)

The Supreme Court struck down Section 66A of the Information Technology Act, 2000, and examined constitutional limitations on online speech regulation.

Relevance: AI-based content moderation and automated restrictions on digital expression must operate consistently with constitutional free-speech protections and applicable legal standards.

4. Anuradha Bhasin v. Union of India (2020)

The Supreme Court addressed restrictions on internet access and emphasized constitutional protections concerning freedom of speech and expression and trade and business.

Relevance: Automated systems used to regulate digital access or impose technological restrictions should remain subject to constitutional standards, proportionality, and judicial scrutiny.

5. Arjun Panditrao Khotkar v. Kailash Kushanrao Gorantyal (2020)

The Supreme Court examined the admissibility of electronic records under Section 65B of the Indian Evidence Act, 1872.

Relevance: Accountability for AI-generated decisions may depend on electronic records, system logs, digital evidence, and the ability to establish the authenticity and reliability of technological outputs.

The judgment provides an important evidentiary foundation for disputes involving computational systems.

6. Jacob Mathew v. State of Punjab (2005)

The Supreme Court examined standards of professional negligence and criminal liability in medical practice.

Relevance: Where AI-assisted medical systems contribute to harmful outcomes, accountability may require examination of professional standards, reasonable care, human supervision, and the responsibilities of medical practitioners.

The judgment does not specifically address AI liability, but its negligence principles provide a useful analytical framework.

7. Donoghue v. Stevenson (1932), House of Lords

This landmark English decision established an important foundation for the modern law of negligence through the neighbour principle.

Relevance: The principle can inform the development of liability standards where developers, manufacturers, or operators fail to exercise reasonable care in designing or deploying AI systems that foreseeably harm others.

5. International legal and regulatory framework

International regulatory developments provide important models for establishing accountability for non-human cognitive systems.

A. European Union Artificial Intelligence Act, 2024

The EU AI Act establishes a risk-based regulatory framework for artificial intelligence.

It distinguishes different categories of AI applications and imposes requirements on high-risk systems concerning risk management, data governance, documentation, transparency, human oversight, accuracy, and cybersecurity.

Its significance lies in recognizing that the level of regulatory responsibility should reflect the potential risks associated with AI applications.

B. OECD AI Principles, 2019

The OECD AI Principles promote trustworthy AI based on:

Inclusive growth and sustainable development.

Human-centred values and fairness.

Transparency and explainability.

Robustness, security, and safety.

Accountability.

These principles emphasize that organizations and individuals responsible for AI systems should remain answerable for their proper functioning.

C. UNESCO Recommendation on the Ethics of Artificial Intelligence, 2021

UNESCO emphasizes human rights, transparency, accountability, fairness, privacy, and human oversight in AI development and deployment.

It provides an international ethical governance framework for managing the social consequences of artificial intelligence.

D. Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law, 2024

This convention establishes principles for ensuring that AI activities remain consistent with human rights, democratic institutions, and the rule of law.

It is particularly relevant to public-sector accountability and safeguards against harmful automated decision-making.

6. Allocation of legal responsibility for non-human cognitive systems

An effective accountability framework must identify the responsibilities of different participants in the AI lifecycle.

A. Developers

Developers may be responsible for failures involving defective design, inadequate testing, foreseeable risks, or failures to implement appropriate safeguards, depending on applicable law.

B. Manufacturers

Manufacturers may face product-related liability where an AI-enabled product contains a legally recognized defect that causes harm.

C. Deploying organizations

Organizations using AI systems must assess whether the technology is suitable for its intended purpose and whether adequate supervision and monitoring mechanisms exist.

D. Human operators

Human operators may have responsibility where they misuse a system, ignore warnings, or fail to exercise reasonable care within their professional duties.

E. Government authorities

Public authorities remain responsible for ensuring that automated administrative processes comply with constitutional requirements, statutory obligations, and principles of natural justice.

Importantly, accountability should be allocated according to actual legal duties, control, conduct, and causation rather than automatically assigning all responsibility to one participant.

7. Major legal and governance challenges

Non-human cognition accountability faces several significant challenges.

Accountability gaps: Complex AI systems may involve multiple developers, data providers, operators, and deploying organizations.

Algorithmic opacity: Some systems make predictions that are difficult to explain or independently reproduce.

Causation difficulties: Establishing the connection between an AI-generated output and the resulting harm may be complicated.

Bias and discrimination: AI systems may reproduce discriminatory patterns present in training datasets.

Data privacy: Extensive data collection may create risks involving unauthorized processing and disclosure.

Autonomous decision-making: Greater technological independence may reduce immediate human involvement in consequential decisions.

Cross-border jurisdiction: AI services may involve developers, infrastructure, and users located in different countries.

Legal uncertainty: Traditional liability frameworks may not adequately address every risk associated with advanced AI systems.

8. Recommendations for developing an accountability framework

A comprehensive governance framework should incorporate the following measures:

Establish clear legal responsibilities for AI developers, manufacturers, operators, and deploying organizations.

Require meaningful human oversight for high-impact automated decisions.

Introduce appropriate transparency and explainability requirements.

Maintain reliable records of consequential AI-generated decisions.

Establish independent auditing and risk-assessment mechanisms.

Protect personal information through appropriate data protection safeguards.

Create accessible complaint and appeal mechanisms for individuals affected by automated decisions.

Develop appropriate civil liability and compensation frameworks for legally recognized AI-related harm.

Require periodic monitoring and testing of high-risk AI systems.

Ensure that public-sector AI decisions remain subject to constitutional and judicial review.

These measures would help prevent technological autonomy from becoming a mechanism for avoiding legal responsibility.

9. Conclusion

Non-Human Cognition Doctrine Accountability represents an important emerging concept in artificial intelligence governance, constitutional law, and technological accountability. Its central objective is to ensure that non-human cognitive systems remain subject to meaningful legal supervision and that their use does not create gaps in responsibility.

The doctrine emphasizes that artificial intelligence may perform increasingly complex cognitive functions, but technological autonomy does not automatically create legal immunity for the individuals and institutions responsible for designing, deploying, and operating these systems.

In India, Articles 14, 19, 21, 32, and 226 provide important constitutional foundations for developing this framework. Judicial decisions such as Justice K.S. Puttaswamy v. Union of India (2017), Shreya Singhal v. Union of India (2015), and Anuradha Bhasin v. Union of India (2020) offer relevant principles concerning privacy, freedom, legality, and judicial review.

Ultimately, effective non-human cognition accountability requires a legal framework in which technological innovation is accompanied by transparency, human responsibility, institutional oversight, and effective remedies. Artificial intelligence may make decisions, but the legal responsibility of the humans and institutions governing its use must remain identifiable and enforceable.