Gujarat Police Launches AI Tool NARIT AI to Boost Drug Case Convictions

The application, named 'NARIT AI' (Narcotics Analysis and RAG-based Investigation Tool), was launched in Gandhinagar as part of the state's zero-tolerance approach to drug-related offences.

It is intended to assist investigating officers in dealing with cases under the Narcotic Drugs and Psychotropic Substances (NDPS) Act, 1985, by providing real-time legal analysis and procedural guidance.

Built on Retrieval Augmented Generation (RAG) technology, the system analyses First Information Reports (FIRs) uploaded by officers and generates detailed reports assessing the strengths and weaknesses of a case.

It compiles relevant legal provisions, case law and investigation procedures, and offers suggestions on legal remedies, compliance requirements and evidence collection.

"This is the first kind of AI-based tool being used by any state police or law enforcement agency in India," officials said, adding that the initiative is aimed at making investigations more effective and strengthening cases during court trials to ensure punishment for the accused.

Deputy Chief Minister Harsh Sanghavi said the state remained committed to dismantling drug networks.

"Gujarat state is determined to eliminate drug-related crimes. We are committed to doing whatever it takes to dismantle this entire network. All necessary arrangements have been made to ensure effective investigation of drug-related crimes and to build a strong legal case against the accused, and with the help of technology, this work will become even stronger," he said.

The application has been developed by the Western Railway Police in the Vadodara division in collaboration with a Mumbai-based AI startup.

It has been implemented under the guidance of Gujarat Director General of Police (DGP), Dr K. L. N. Rao and Surat City Police Commissioner, Anupam Singh Gehlot, with Western Railway Vadodara SP, Abhay Soni leading the initiative.

Soni said the tool addresses a long-standing challenge in narcotics investigations, where procedural lapses have led to acquittals despite strong evidence.

"Earlier, investigations were carried out only by specially trained police staff, which led to workload and administrative delays. This tool will act as a 'force multiplier' by using advanced RAG-based AI technology and will empower every investigating officer to conduct proper investigations as per rules, thereby improving the conviction rate in NDPS cases. As this application is trained on High Court and Supreme Court landmark cases, bare acts, and government guidelines, the possibility of 'hallucination' is very low, and it assists investigating officers in conducting scientific investigations," he said.

Police officials said procedural compliance remains one of the biggest challenges in drug-related cases.

The application is designed to help officers avoid such lapses by identifying weaknesses in FIRs, advising on required steps, outlining what should and should not be done during investigations, and providing a checklist of evidence.

The generated reports can also be used while taking final decisions, improving the quality and consistency of investigations.

The system has been trained on landmark judgments of High Courts and the Supreme Court, as well as on legal frameworks including the NDPS Act, 1985, and the bare acts of BNS, BNSS and BSA, along with government guidelines and circulars.

It operates on a closed database and does not rely on open internet searches, a feature intended to reduce the likelihood of inaccurate outputs.

In line with recent guidelines of the Gujarat High Court on the use of artificial intelligence, the application is classified as a private AI system developed specifically for Gujarat Police and is not available to the general public.

In addition to supporting investigations, the tool can assist during court proceedings by anticipating possible defence arguments and suggesting responses based on established case law from High Courts and the Supreme Court, helping to present cases more effectively during trial.

- IANS