DOES AI-BASED PREDICTIVE POLICING IN PREVENTIVE DETENTION MEET INDIA'S CONSTITUTIONAL PROPORTIONALITY STANDARD?

Artificial Intelligence tools are increasingly being incorporated into the Indian police forces, from Delhi Police's Crime Mapping, Analytics and Predictive System (CMAPS) to Bengaluru's C4i platform and Akola's Project Trinetra. These algorithms serve the purpose of assigning machine-generated risk scores to repeat offenders, or identify areas of possible crime occurrence, to predict crime before it occurs. .As of now these systems are not considered formal triggers for preventive detention. But the path from area-based hotspot mapping, to individual risk assessment or scoring makes that step seem likely. This paper takes up an analysis regarding the prospective proportionality of this method. Would a regime using AI-generated risk scores formally to inform decisions to invoke preventive detention, under statutes such as the National Security Act, 1980, state Public Safety Acts, or Section 151 of the Code of Criminal Procedure, 1973, satisfy the four-fold proportionality standard as conceptualized in Justice K.S. Puttaswamy (Retd.) v. Union of India. It also looks into the requirement for concrete safeguards to exist before such a system could be constitutionally defensible.