Introduction

Risk prediction for youth crime governance refers to the use of statistical tools, artificial intelligence (AI), social information, or professional assessments to estimate whether a young person may be at risk of offending or reoffending. Such systems may be used by police, juvenile justice authorities, probation officers, or social services to identify support needs and plan rehabilitation.

The issue is legally sensitive because predicting future criminal behaviour is not the same as proving that a person has committed an offence. A young person may be classified as “high risk” because of family circumstances, school absence, poverty, neighbourhood conditions, or previous police contact. These factors may reflect social disadvantage rather than criminal intent.

In India, there is no specific law authorising an AI algorithm to determine a child’s future criminality. The Juvenile Justice (Care and Protection of Children) Act, 2015, instead emphasises dignity, rehabilitation, privacy, the best interests of the child, and individualised assessment. Therefore, risk prediction must remain a carefully controlled aid—not a substitute for legal judgment or a basis for automatically imposing punishment.

Legal Framework and Case Laws

1. Barun Chandra Thakur v. Master Bholu, (2022) 2 SCC 401

The Supreme Court examined preliminary assessment under Section 15 of the Juvenile Justice Act, 2015, for a child aged 16 or above alleged to have committed a heinous offence. The assessment concerns the child’s mental and physical capacity, ability to understand consequences, and circumstances of the alleged offence.

The Court stressed that this is a delicate, individualised task requiring expertise. It noted that cognitive and emotional development continues into early adulthood and called for appropriate guidance for Juvenile Justice Boards. This case is important because a predictive score cannot replace the statutory assessment or determine that a child should be tried as an adult.

2022 Supreme(SC) 568+1

2. K.S. Puttaswamy v. Union of India, (2017) 10 SCC 1

The Supreme Court recognised privacy as a fundamental right under Article 21, including informational privacy and personal autonomy. Applying this principle, collecting and combining a child’s school, family, location, or police data to predict future offending may interfere with privacy.

Such processing must have a lawful basis, pursue a legitimate aim, and satisfy proportionality. Sensitive information should not be collected or retained merely because it might improve a prediction.

Sci API

3. In Re: Right to Privacy of Adolescents, 2025 INSC 778

The Supreme Court considered the privacy and dignity of adolescents in the context of criminal proceedings. The judgment reinforces the need to protect children from stigma and unnecessary exposure of sensitive personal information. It is relevant by analogy to predictive profiling, which may create lasting labels about a child’s supposed future behaviour.

LawDistill

Principles for Responsible Governance

Risk prediction should follow these safeguards:

No automatic adverse action: A risk score must not alone justify detention, surveillance, or transfer to adult trial.

Human oversight: Qualified professionals and legal authorities must independently assess each case.

Bias prevention: Data and models must be tested for discriminatory effects.

Privacy and confidentiality: Children’s records must be protected and used only for lawful, necessary purposes.

Rehabilitation priority: Predictions should guide access to counselling, education, and support—not stigmatise children.

Conclusion

Risk prediction for youth crime governance must be governed by child rights, constitutional privacy, and individualised justice. Barun Chandra Thakur establishes the importance of expert, child-specific assessment, while Puttaswamy provides the constitutional basis for limiting intrusive data use. Predictive tools may assist rehabilitation planning, but they cannot establish guilt or replace statutory decision-making. A lawful system must be transparent, proportionate, independently reviewed, and directed toward rehabilitation rather than punishment based on predicted behaviour.