The Human Mind on Trial: Constitutional Limits of Artificial Intelligence in Criminal Investigation

Abstract:

Artificial Intelligence (AI) is reshaping criminal investigations by enabling law enforcement agencies to analyse vast amounts of data, identify behavioural patterns, employ facial recognition systems, conduct predictive policing, and utilise algorithmic decision-making.
While these technologies promise greater efficiency in crime detection and prevention, they also raise profound constitutional concerns regarding privacy, personal autonomy, due process, and the privilege against self-incrimination. Indian criminal jurisprudence has traditionally focused on regulating coercive investigative techniques such as narco-analysis, polygraph examinations, and Brain Electrical Activation Profile (BEAP) tests. However, the emergence of AI-assisted investigative tools presents a fundamentally different challenge: rather than compelling individuals to disclose information, these technologies increasingly infer thoughts, intentions, emotions, and behavioural tendencies through algorithmic analysis.
This paper argues that the constitutional reasoning adopted by the Supreme Court in Selvi v. State of Karnataka extends beyond the prohibition of involuntary scientific techniques and provides a broader framework for protecting mental autonomy against technologically mediated intrusions. It contends that the principles underlying Articles 20(3) and 21 of the Constitution must evolve to address contemporary forms of cognitive surveillance, including facial emotion recognition, behavioural analytics, predictive policing, biometric intelligence, and emerging neurotechnologies. By revisiting Selvi through the lens of artificial intelligence, the paper examines whether existing constitutional safeguards remain adequate in an era where technology can generate incriminating inferences without obtaining a conventional confession.
Adopting a doctrinal and comparative research methodology, the study analyses constitutional provisions, judicial precedents, statutory frameworks, and international developments relating to AI governance and criminal justice. It compares India’s legal position with evolving approaches in the European Union, the United Kingdom, the United States, UNESCO, and the OECD to identify emerging principles concerning transparency, accountability, and the protection of mental privacy.

The paper concludes that constitutional protections against self-incrimination and arbitrary State action should not be confined to traditional investigative methods. Instead, they must be interpreted dynamically to preserve human dignity, cognitive liberty, and fair trial guarantees in the age of artificial intelligence. It proposes a principled constitutional and regulatory framework that balances technological innovation with the enduring values of liberty, privacy, and the rule of law, ensuring that advances in criminal investigation do not undermine the fundamental rights that define a constitutional democracy.

Keywords: Artificial Intelligence; Criminal Investigations; Constitutional Law; Article 20(3); Article 21; Self-Incrimination; Selvi v. State of Karnataka; Cognitive Surveillance; Mental Privacy; Cognitive Liberty; Predictive Policing; Human Rights; Criminal Justice; Algorithmic Accountability; Due Process.

Introduction

The integration of Artificial Intelligence (AI) into criminal justice has transformed the manner in which crimes are detected, investigated, and prosecuted. Technologies that once belonged to the realm of science fiction—facial recognition systems, predictive policing algorithms, biometric identification, behavioural analytics, automated risk assessment, and AI-assisted digital forensics—are increasingly becoming part of modern law enforcement. These innovations promise greater efficiency, faster investigations, and enhanced public safety by enabling authorities to process vast quantities of information beyond human capacity. However, the growing reliance on intelligent technologies has also generated complex constitutional and ethical questions concerning individual liberty, privacy, human dignity, and the limits of State power.

Criminal investigations have historically depended upon physical evidence, eyewitness testimony, documentary records, and voluntary statements made by accused persons. Constitutional democracies, including India, have long recognised that while the State possesses the authority to investigate crime, such authority cannot be exercised without constitutional restraint. The criminal justice system therefore seeks to maintain a delicate balance between two competing objectives: protecting society through effective law enforcement while safeguarding the fundamental rights of individuals subjected to investigation and prosecution. This balance becomes increasingly fragile when technological innovations enable the State to acquire information in ways that traditional legal doctrines never anticipated.

The rapid advancement of AI has fundamentally altered this landscape. Unlike conventional investigative tools, AI systems are capable of identifying patterns, predicting behaviour, recognising emotions, analysing speech, interpreting digital footprints, and generating probabilistic conclusions regarding an individual’s actions or intentions. These technologies often function through algorithmic inference rather than direct evidence. Instead of compelling an accused person to make a statement, AI may derive conclusions from behavioural data, biometric information, communication patterns, location history, or facial expressions. Such capabilities challenge conventional understandings of evidence, voluntariness, and self-incrimination, compelling legal systems to reconsider the constitutional principles governing criminal investigations.

The constitutional implications of AI are particularly significant in India, where the protection of personal liberty occupies a central position within constitutional jurisprudence. Articles 20(3) and 21 of the Constitution of India embody two of the most important safeguards against arbitrary State action. Article 20(3) guarantees that no person accused of an offence shall be compelled to be a witness against himself, while Article 21 protects life and personal liberty except according to a procedure established by law that is just, fair, and reasonable. Through judicial interpretation, these provisions have evolved beyond their literal text to encompass broader principles of dignity, privacy, autonomy, and procedural fairness. Consequently, constitutional scrutiny must extend not only to traditional forms of coercion but also to emerging technologies capable of indirectly obtaining or generating incriminating information.

A significant milestone in this constitutional development was the Supreme Court’s decision in Selvi v. State of Karnataka. The Court held that the involuntary administration of narco-analysis, polygraph examinations, and Brain Electrical Activation Profile (BEAP) tests violated the constitutional guarantee against self-incrimination and infringed the right to personal liberty and mental privacy. The judgment reaffirmed that constitutional protections extend beyond physical coercion and encompass respect for an individual’s cognitive autonomy, dignity, and freedom of choice during criminal investigations. Although the decision was delivered in the context of specific forensic techniques, its reasoning reflects broader constitutional concerns regarding technological intrusion into the human mind.

More than a decade after Selvi, the technological environment has undergone a remarkable transformation. Artificial intelligence now performs investigative functions that were unimaginable when the judgment was delivered. Facial recognition systems can identify individuals from surveillance footage within seconds. Predictive policing algorithms analyse historical crime data to identify potential offenders or locations associated with future criminal activity. Behavioural analytics evaluate patterns of online conduct, digital interactions, and communication habits. Emotion recognition technologies claim to infer psychological states from facial movements or vocal characteristics, while advances in neuroscience and brain-computer interfaces have revived debates concerning the legal boundaries of cognitive surveillance. Although these technologies differ significantly from narco-analysis or polygraph examinations, they raise a common constitutional concern: the increasing ability of the State to infer information about an individual’s thoughts, intentions, or mental processes without conventional testimony.

This development exposes an important gap in contemporary constitutional discourse. Existing legal debates concerning AI in criminal justice frequently focus on issues such as algorithmic bias, reliability, transparency, data protection, and privacy. While these concerns remain important, comparatively little attention has been devoted to the relationship between AI-driven investigative technologies and the constitutional privilege against self-incrimination. The legal question is no longer confined to whether the State can compel an accused person to speak; it now extends to whether the State may constitutionally rely upon intelligent technologies that generate incriminating inferences without requiring any verbal disclosure. If technology can reconstruct aspects of an individual’s mental state through data-driven analysis, does this represent a new form of compelled testimony, or does it fall outside the traditional scope of constitutional protection? Existing jurisprudence offers no definitive answer.

This paper argues that the constitutional significance of Selvi extends far beyond its immediate prohibition of involuntary scientific techniques. Rather than viewing the judgment as a narrow decision concerning narco-analysis, it proposes that Selvi should be understood as establishing a broader constitutional principle: the State must not employ technology in a manner that undermines an individual’s mental autonomy, cognitive freedom, or the constitutional privilege against self-incrimination. The paper contends that this principle acquires renewed importance in the age of AI, where investigative technologies increasingly operate by producing algorithmic inferences rather than obtaining direct confessions. Accordingly, constitutional protections should evolve to address emerging forms of cognitive surveillance while preserving the balance between effective criminal investigation and individual liberty.

To develop this argument, the paper adopts a doctrinal and comparative approach. It analyses constitutional provisions, judicial precedents, statutory frameworks, and scholarly literature concerning AI, criminal investigations, privacy, and self-incrimination. It further examines comparative developments in jurisdictions such as the European Union, the United Kingdom, and the United States, alongside international guidance issued by organisations including UNESCO and the OECD, to identify emerging standards governing the responsible use of AI within criminal justice systems. These comparative perspectives provide valuable insight into the challenges of regulating rapidly evolving investigative technologies while respecting constitutional guarantees.

Ultimately, this paper seeks to contribute to the emerging discourse on AI and constitutional criminal procedure by advancing a central proposition: constitutional safeguards must evolve alongside technological innovation. The privilege against self-incrimination cannot remain confined to protecting individuals against compelled verbal testimony if intelligent technologies increasingly enable the State to infer information that individuals themselves are constitutionally entitled not to disclose. The enduring relevance of Selvi therefore lies not merely in its rejection of involuntary scientific techniques, but in its broader affirmation that the human mind remains a constitutionally protected space. As artificial intelligence becomes an increasingly influential participant in criminal investigations, preserving that protected space will be essential to ensuring that technological progress strengthens, rather than diminishes, the rule of law, human dignity, and the constitutional promise of justice.

From Traditional Policing to Intelligent Investigations: The Evolution of Artificial Intelligence in Criminal Justice

Criminal investigations have always evolved alongside technological innovation. From fingerprint analysis and DNA profiling to digital forensics and cyber investigations, each advancement has expanded the investigative capabilities of law enforcement while simultaneously challenging established legal principles. Artificial Intelligence represents the latest and perhaps the most transformative stage of this evolution.

This chapter traces the development of AI in criminal investigations, examining how technologies such as facial recognition systems, predictive policing, automated risk assessment, behavioural analytics, voice recognition, biometric identification, and digital forensic tools have become integral to contemporary policing. While these technologies promise greater efficiency and accuracy, they also introduce new concerns regarding algorithmic bias, transparency, accountability, and the potential erosion of individual liberties. The chapter argues that AI has fundamentally shifted criminal investigations from evidence-based reconstruction to data-driven prediction, necessitating a corresponding evolution in constitutional safeguards.

Constitutional Boundaries of Criminal Investigations: Articles 20(3), 21, and the Protection of Human Dignity

The Indian Constitution recognises that the pursuit of justice cannot come at the cost of fundamental rights. Articles 20(3) and 21 establish constitutional limits on investigative powers by protecting individuals against compelled self-incrimination, arbitrary State action, and violations of personal liberty. Judicial interpretation has further expanded these guarantees to include dignity, privacy, fairness, and procedural due process.

This chapter examines the constitutional foundations governing criminal investigations, analysing how landmark decisions have transformed these provisions into living guarantees capable of responding to evolving State power. It argues that constitutional rights are not static protections against historical forms of coercion but dynamic principles that must remain effective in the face of emerging technological challenges.

Re-reading Selvi v. State of Karnataka: Beyond Narco-Analysis Towards Constitutional Protection of the Human Mind

Selvi v. State of Karnataka is often remembered as the judgment that

declared the involuntary administration of narco-analysis, polygraph examinations, and Brain Electrical Activation Profile (BEAP) tests unconstitutional. While this understanding is correct, it captures only part of the judgment’s constitutional significance. A closer reading reveals that the Supreme Court was addressing a far more fundamental concern: the constitutional limits of State power when technology is used to penetrate the human mind during criminal investigations.

Before Selvi, scientific techniques were increasingly perceived as innovative tools capable of accelerating criminal investigations and uncovering information that conventional methods could not obtain. Supporters argued that these techniques could assist in solving complex crimes where direct evidence was scarce. However, their growing use also raised profound constitutional questions. Could the State compel an individual to undergo procedures designed to extract information from their mind? Would technological efficiency justify limiting the constitutional guarantees of dignity, autonomy, and the privilege against self-incrimination?

In its landmark judgment, the Supreme Court answered these questions decisively. The Court held that involuntary narco-analysis, polygraph tests, and BEAP examinations violate Article 20(3) of the Constitution because they compel an accused person to become a witness against themselves. Equally important, the Court recognised that such techniques infringe Article 21 by interfering with an individual’s mental autonomy, personal liberty, and dignity. The judgment affirmed that constitutional protections extend beyond protection from physical coercion; they also safeguard the freedom of an individual to control their own thoughts, choices, and cognitive processes.

Although the Court’s reasoning emerged from the context of specific forensic techniques, its constitutional philosophy reaches much further. The judgment acknowledged that the human mind occupies a unique position within constitutional jurisprudence. Unlike fingerprints, blood samples, handwriting specimens, or other forms of physical evidence, an individual’s thoughts, memories, emotions, and cognitive responses represent deeply personal aspects of human autonomy. The Constitution therefore imposes strict limitations on any investigative practice that seeks to access or manipulate these mental processes without genuine consent.

This paper argues that the true legacy of Selvi does not lie merely in prohibiting particular scientific tests. Rather, the judgment established a broader constitutional principle: the State cannot employ technology to bypass an individual’s freedom to think, choose, and remain silent. This principle is increasingly relevant in an era where artificial intelligence has transformed the nature of criminal investigations.

Unlike narco-analysis, contemporary AI systems rarely seek direct verbal confessions. Instead, they analyse facial expressions, voice patterns, behavioural data, biometric information, online activities, and digital footprints to generate probabilistic assessments about an individual’s intentions, credibility, emotions, or future conduct. These technologies often operate without physical force or chemical intervention, yet they may reveal information that an individual has consciously chosen not to disclose. Consequently, the constitutional challenge has shifted from compelled testimony to algorithmic inference.

This shift raises a critical legal question: if the Constitution prohibits the State from chemically compelling an individual to reveal information, should it also prohibit technologies that algorithmically infer substantially similar information from behavioural or biometric data? Although Selvi did not confront artificial intelligence, its underlying constitutional reasoning provides an important framework for addressing this question.

The significance of Selvi becomes even clearer when viewed alongside the Supreme Court’s later privacy jurisprudence. The recognition of privacy, dignity, and personal autonomy as essential constitutional values reinforces the idea that the human mind cannot become an unrestricted source of investigative data merely because technology has made such access possible. Constitutional rights must remain effective despite technological progress; otherwise, advances in investigative techniques would gradually erode protections that the Constitution was designed to preserve.

At the same time, this chapter does not suggest that every use of AI in criminal investigations is unconstitutional. Artificial intelligence can significantly improve forensic analysis, assist in identifying suspects, detect financial fraud, and process large volumes of digital evidence. These legitimate objectives cannot be ignored in an increasingly complex criminal justice system. The constitutional challenge is therefore not whether AI should be prohibited, but how its use should be regulated so that technological efficiency does not undermine the guarantees of fairness, dignity, autonomy, and due process.

Re-reading Selvi through this broader constitutional lens reveals that it was not merely a judgment about narco-analysis. It was an early articulation of a constitutional principle that remains highly relevant today: technological innovation cannot justify unrestricted intrusion into the human mind. As investigative practices evolve from extracting confessions to generating algorithmic inferences, the constitutional values recognised in Selvi must evolve as well. Its enduring contribution lies in reminding us that the effectiveness of criminal investigations must always remain subordinate to the constitutional commitment to human dignity, individual autonomy, and the rule of law.

Conclusion

Artificial Intelligence has ushered in a new era of criminal investigations, fundamentally transforming the way law enforcement agencies collect, analyse, and interpret evidence. Technologies such as facial recognition, predictive policing, behavioural analytics, biometric identification, and algorithmic decision-making have enhanced investigative efficiency and expanded the State’s capacity to prevent and detect crime. Yet, these developments have also challenged long-established constitutional principles that were designed to protect individuals from arbitrary exercises of State power. The increasing ability of AI to infer information about a person’s thoughts, intentions, emotions, and behaviour raises questions that traditional criminal procedure was never designed to answer.

This paper has argued that the constitutional significance of Selvi v. State of Karnataka extends well beyond its immediate prohibition of involuntary narco-analysis, polygraph examinations, and Brain Electrical Activation Profile (BEAP) tests. Although the judgment addressed specific forensic techniques, its enduring contribution lies in affirming a broader constitutional principle: the State cannot employ technology in a manner that compromises an individual’s mental autonomy, dignity, or the privilege against self-incrimination. When read in the context of contemporary technological developments, Selvi emerges not merely as a case on scientific evidence but as a foundational precedent for regulating technologically assisted criminal investigations.

The transition from compelled confessions to algorithmic inferences marks a significant shift in the constitutional landscape. Modern AI systems do not necessarily compel an accused person to speak; instead, they analyse behavioural patterns, digital footprints, biometric data, and other forms of personal information to generate conclusions that may influence criminal investigations and judicial outcomes. This development creates a constitutional dilemma. If technology can reconstruct aspects of an individual’s mental processes without obtaining a conventional confession, the privilege against self-incrimination risks becoming increasingly ineffective unless interpreted in light of these new realities. Constitutional protections must therefore evolve to address not only direct compulsion but also technologically mediated forms of cognitive intrusion.

Comparative developments across jurisdictions demonstrate that this concern is neither theoretical nor confined to India. The European Union has introduced regulatory safeguards for high-risk AI systems, while courts and policymakers in the United States and the United Kingdom continue to debate the admissibility, transparency, and accountability of algorithmic evidence. International organisations such as UNESCO and the OECD have similarly emphasised that AI governance must remain grounded in respect for human rights, transparency, fairness, and democratic values. Although these initiatives do not provide a complete solution, they reflect an emerging global consensus that technological innovation must operate within constitutional and ethical boundaries.

The analysis undertaken in this paper also reveals important gaps in the Indian legal framework. Existing constitutional and statutory protections primarily regulate conventional investigative techniques and provide limited guidance on the use of AI-assisted tools in criminal investigations. There is no comprehensive legislative framework governing facial recognition systems, predictive policing, behavioural analytics, or algorithmic decision-making in law enforcement. Questions concerning algorithmic bias, explainability, judicial oversight, evidentiary reliability, and accountability remain largely unresolved. These regulatory deficiencies create uncertainty for investigators, courts, and individuals whose rights may be affected by AI-driven investigative practices.

Accordingly, this paper has proposed that India’s constitutional and legislative framework should evolve in a manner that preserves both technological innovation and fundamental rights. The principles underlying Articles 20(3) and 21 should be interpreted dynamically to encompass emerging forms of cognitive surveillance and AI-enabled investigative techniques. Legislative reforms should establish clear standards governing the deployment of AI in criminal investigations, including requirements relating to legality, necessity, proportionality, transparency, human oversight, independent auditing, and judicial review. Such safeguards would not impede legitimate law enforcement objectives; rather, they would strengthen public confidence by ensuring that technological advancements remain accountable to constitutional values.

Ultimately, the question confronting modern criminal justice is not whether Artificial Intelligence should become part of criminal investigations—it already has. The real constitutional challenge is determining the limits within which these technologies may operate without eroding the liberties they are intended to protect. The Constitution does not resist technological progress; it insists that progress remain consistent with the principles of justice, dignity, equality, and the rule of law.

As criminal investigations become increasingly dependent on intelligent technologies, the true legacy of Selvi lies in its reminder that the human mind cannot become an unrestricted object of State scrutiny simply because technology makes such scrutiny possible. Constitutional democracy demands that scientific innovation serve justice without compromising individual freedom. In this sense, protecting the human mind is not merely a matter of privacy—it is an affirmation of the constitutional promise that every person, regardless of accusation, retains the dignity, autonomy, and liberty that define the rule of law.

Cases

• Selvi v. State of Karnataka, (2010) 7 SCC 263.
• Justice K.S. Puttaswamy (Retd.) v. Union of India, (2017) 10 SCC 1.
• Maneka Gandhi v. Union of India, (1978) 1 SCC 248.
• Nandini Satpathy v. P.L. Dani, (1978) 2 SCC 424.
• State of Bombay v. Kathi Kalu Oghad, AIR 1961 SC 1808.
• D.K. Basu v. State of West Bengal, (1997) 1 SCC 416.

Statutes

• Constitution of India.
• Bharatiya Nagarik Suraksha Sanhita, 2023.
• Bharatiya Sakshya Adhiniyam, 2023.
• Information Technology Act, 2000.
• Digital Personal Data Protection Act, 2023.

Books

• M.P. Jain, Indian Constitutional Law.
• H.M. Seervai, Constitutional Law of India.
• V.N. Shukla, Constitution of India.
• K.N. Chandrasekharan Pillai, R.V. Kelkar’s Criminal Procedure.
• Frank Pasquale, The Black Box Society.
• Cathy O’Neil, Weapons of Math Destruction.

International Documents

• UNESCO, Recommendation on the Ethics of Artificial Intelligence (2021).
• OECD, OECD Principles on Artificial Intelligence (2019).
• European Union, Artificial Intelligence Act.
• European Union, General Data Protection Regulation (GDPR).
• Council of Europe, Framework Convention on Artificial Intelligence, Human Rights, Democracy and the Rule of Law (2024).

Author: Sanskriti Pathak