AI is revolutionizing India’s criminal justice system (CJS) by enhancing efficiency in investigation, adjudication, surveillance, and evidence analysis under Industry 4.0.

AI in Criminal Justice

• Artificial Intelligence (AI) is transforming justice delivery across investigation, adjudication, surveillance, and evidence analysis. 
• In India, the use of AI-based tools like CCTNS, ICJS, AFRS, and SUPACE is expanding rapidly, aiming to enhance speed, efficiency, and predictability. 
• However, this raises serious concerns about bias, fairness, constitutional morality, and data privacy.

Areas of AI Use in India’s Criminal Justice System

• Law Enforcement and Policing
  • Predictive Policing: AI tools analyze historical crime data to predict potential crime hotspots.
    • Example: Telangana CMAPS (Crime Mapping, Analytics and Predictive System) uses AI for proactive patrolling.
  • Facial Recognition Systems: Used to match CCTV footage with police databases.
    • Helps in identifying suspects, locating missing persons, and crowd surveillance.
  • Surveillance and Video Analytics: AI-enabled CCTV systems detect abnormal movements, track suspects, and alert control rooms.
    • Helps reduce manpower burden on monitoring teams.
• Investigation and Forensics
  • Digital Evidence Analysis: AI helps process large volumes of digital evidence (e.g., email, chats, CCTV).
    • Enhances speed and accuracy of forensic investigations.
  • Pattern Recognition: Identifies criminal modus operandi by analyzing behaviour across cases.
  • Forensic Science: AI used in voice matching, DNA analysis, and fingerprint recognition.
• Case Management & Legal Research:
  • SUPACE (Supreme Court Portal for Assistance in Court’s Efficiency): An AI tool launched in 2021 to help judges go through large volumes of case data and precedents.
    • AI does not make judgments—used only for legal research support.
  • E-Courts and Smart Filing: Courts use AI for scanning, sorting, and tagging digital case files.
    • Speeds up case processing and identifies missing or defective pleadings.
• Prison & Rehabilitation
  • AI for Prisoner Risk Assessment: Analyzes recidivism risks (similar to USA’s COMPAS, which was found racially biased).
  • e-Prisons System: Tracks inmate records, parole eligibility, and rehabilitation programs.

Government Initiatives on AI in Criminal Justice

• Crime and Criminal Tracking Network and Systems (CCTNS): Launched in 2009 under National e-Governance Plan.
  • Nodal Agency: NCRB, Ministry of Home Affairs.
  • Objective: Create a nationwide searchable digital database of crimes and criminals.
  • AI Relevance: Enables predictive policing, behavioral pattern analysis, and cross-jurisdictional data linking.
• Inter-operable Criminal Justice System (ICJS): Launched: Phase 1 (2014), Phase 2 approved in 2022 with ₹3,700 crore allocation.
  • Objective: Seamless data exchange between police, prisons, courts, prosecution, and forensics.
  • AI Relevance: AI can assist in linking cases, matching criminal profiles, and tracking case histories.
• National Automated Facial Recognition System (AFRS): Developed By NCRB.
• • Purpose: Facial recognition of suspects, missing persons, and criminals using CCTV inputs and crime databases.

State-Level AI Policing Initiatives

Benefits of AI in Criminal Justice

• Enhanced Efficiency and Speed: Automation of routine tasks such as case tagging, documentation, FIR registration, and record-keeping.
  • Example: AI-based tools can sort through thousands of pages of evidence and judgments in seconds.
• Improved Investigative Accuracy: Facial recognition systems (AFRS) help identify suspects or missing persons from vast video surveillance data.
  • AI in digital forensics can scan emails, phone logs, and social media for hidden patterns.
  • Enhances link analysis between suspects, events, and criminal networks.
• Data-Driven Decision-Making: AI tools provide objective inputs for crime forecasting, bail, sentencing risk assessments, and parole decisions.
  • Ensures consistency and evidence-based policing, minimizing arbitrary judgments.
  • Tools like CCTNS and ICJS enable seamless data sharing among: Police → Courts → Prisons → Forensics → Prosecution.
  • AI can analyze large cross-institutional datasets to detect case overlaps, missed links, and procedural delays.
• Preventive Policing and Crime Deterrence: Predictive policing can preempt crimes by identifying potential offenders or vulnerable locations.
  • Reduces human bias in routine surveillance and ensures round-the-clock automated vigilance.
• Empowerment of Law Enforcement Personnel: Reduces workload and cognitive fatigue through automation.
  • Frees up police officers for community engagement and strategic fieldwork rather than administrative tasks.
• Policy Evaluation and Systemic Insights: Analysis of long-term crime data can inform criminal justice reforms, budgeting, and resource allocation.
  • Helps governments identify structural issues, like delays in trials, or underreporting trends.

Ethical Concerns of AI in Criminal Justice

• Algorithmic Bias and Systemic Discrimination: AI tools trained on historical datasets reinforce biases related to caste, religion, gender, and region.
  • Example: Dalits, Muslims, and Adivasis disproportionately represented in NCRB prison data .
  • Facial recognition systems have error rates up to 34.7% for dark-skinned women .
  • Structural injustice is deepened, violating Article 14 and 15.
• Digital Exclusion of Marginalized Groups: AI tools rely on digital footprints, but only 31% of rural citizens and 25% of women have internet access.
  • Dalits and Adivasis are severely underrepresented in datasets, risking algorithmic exclusion.
  • Reinforces structural inequality and violates inclusive governance principles.
• Lack of Explainability and Accountability: AI systems used in policing and judicial assistance are often “black boxes”, offering no reasoning behind outcomes.
  • No statutory or institutional mechanism exists to audit algorithmic decisions.
  • Undermines due process, right to reasoned decision, and legal transparency.
• Mass Surveillance and Privacy Intrusion: AI-enabled systems like AFRS can be used for mass surveillance without consent or legal limits.
  • Violates the Right to Privacy upheld in K.S. Puttaswamy v. Union of India (2017).
  • Infringes on individual autonomy, freedom of movement, and digital dignity.
• Erosion of Human Judgment and Discretion: AI is increasingly used to guide judicial and police decisions, but lacks contextual understanding, empathy, or nuanced moral reasoning.
  • Even tools like SUPACE used by the Supreme Court have been limited to research precisely due to these ethical limits.
  • Reliance on data models risks outsourcing moral judgment to machines.
  • Reduces human agency and may lead to dehumanised justice delivery.
• No Legal or Ethical Oversight Framework: India has no AI-specific legislation to regulate its deployment in law enforcement.
• • Ethical concerns like consent, audit, fairness, and redress remain unaddressed.

Challenges for AI in Criminal Justice 

• Fragmented and Non-Uniform Data Ecosystem: Criminal records in India are often non-digitised, inconsistent, or in regional languages, leading to low-quality inputs for AI systems.
  • CCTNS and ICJS still suffer from data backlog and legacy formats.
• Data Security and Misuse: Centralized systems like CCTNS and ICJS collect vast sensitive personal data.
  • Risk of data breaches, hacking, and manipulation in the absence of robust encryption or legal redress.
• Lack of Legal Backing – Constitutional Morality Deficit: The deployment of AI tools in the criminal justice system is not backed by a statutory framework, violating the “procedure established by law” requirement under Article 21.
  • Bypassing legislative scrutiny contradicts the spirit of constitutional governance and rule of law.
• Lack of Skilled Manpower in Law Enforcement: Police officers and judicial staff often lack training in handling AI systems, leading to misuse or overreliance on automated tools.
  • Capacity-building is weak; most officers treat AI as a mystical black box.
• Poor Infrastructure in Lower Courts and Rural Areas: Digital infrastructure like stable internet, secure servers, and device availability is missing in many districts, making AI deployment uneven.
  • AI remains urban-centric; over 70% of district courts lack advanced digital setups.
• Inter-departmental Coordination Gaps: The five pillars of justice — police, courts, prosecution, prisons, forensics — often work in silos, slowing down interoperable AI deployment.
  • ICJS Phase II is still struggling with data standardization and integration.
• Others: 
  • Absence of Pilot Testing and Phased Rollouts.
  • Judges and magistrates are often skeptical of algorithmic tools, preferring manual processes.

Global example of AI in Criminal Justice

Way Forward: AI in Criminal Justice in India

• Enact a Dedicated Legal Framework for AI Use: Introduce legislation that governs the deployment, oversight, audit, and grievance redress of AI in the criminal justice system.
  • Should include provisions on transparency, explainability, accountability, and rights-based safeguards.
• Establish Independent AI Ethics Oversight Authority: A statutory body should certify, audit, and evaluate algorithms used in surveillance, sentencing, or policing.
  • Must include representation from civil society, marginalized groups, and legal experts.
• Ensure Inclusive and Representative Datasets: AI systems must be trained on datasets that reflect India’s diversity — across caste, region, gender, language, and socio-economic status.
  • The government should mandate bias impact assessments and require disaggregated data analysis.
• Develop Explainable and Interpretable AI Systems: Ensure that algorithms provide clear, traceable, and legally admissible reasoning for their decisions.
  • Necessary especially in bail, sentencing, and risk assessments.
• Institutional Capacity Building and Sensitization: Train judges, police, and correctional officers in the ethical, technical, and legal dimensions of AI use.
  • Emphasize bias detection, data ethics, and the limits of AI interpretation.
• Strengthen Data Protection and Privacy Law: Expedite passage of a comprehensive Digital Personal Data Protection Act with specific safeguards for law enforcement AI use.
  • Ensure citizen control over personal data and legal recourse in case of misuse.
• Pilot Programs with Sunset Clauses and Review Mechanisms: Any new AI deployment should begin as pilots with sunset clauses, followed by independent evaluation before full-scale adoption.
  • Focus should be on transparency, accountability, and redressal effectiveness.

Conclusion

AI enhances India’s criminal justice system by improving efficiency and integration, but risks deepening biases and privacy violations without regulation. Ethical, inclusive, and legally governed AI adoption is essential for equitable justice.

DOWNLOAD PDF