In a significant move to curb escalating GST input tax credit (ITC) frauds, the state government is planning to deploy an advanced artificial intelligence (AI) and machine learning (ML)-driven system to detect suspicious transactions and entities. The initiative comes amid mounting losses to the state exchequer, with ITC scams reportedly costing thousands of crores of rupees annually.
Tie-Up with Bhaskaracharya National Institute for Space Applications and Geo-informatics
To implement this technology-driven solution, the commercial tax division of the state’s finance department has partnered with the Bhaskaracharya National Institute for Space Applications and Geo-informatics (BISAG-N). The institute will develop a comprehensive AI-powered framework capable of analysing vast volumes of GST transaction data to identify patterns indicative of fraud.
Massive Volume of Transactions Drives Need for AI
Government sources highlighted that nearly 29 crore ITC-related transactions are recorded in the state annually. Given the sheer scale, manual verification of transactions has become impractical.
Officials noted that the adoption of AI and ML technologies is now essential to efficiently sift through such massive datasets and pinpoint irregularities. The system will analyse transactional behaviour, detect anomalies, and flag entities that exhibit characteristics commonly associated with fraudulent activity.
AI Model Under Development and Testing
Sources revealed that BISAG-N has already developed a large language model (LLM), which is currently undergoing testing and refinement. Once fully operational, the system will process all ITC-related transactions and generate alerts for potentially fraudulent claims and suspicious taxpayers.
The model is expected to continuously improve through fine-tuning, enhancing its accuracy in identifying complex fraud patterns over time.
Enhanced Enforcement and Accountability
Beyond identifying fraudulent taxpayers, the AI system is also designed to strengthen internal accountability. Officials indicated that the model will flag irregular conduct by government personnel as well, addressing instances where insiders may have been complicit in ITC-related frauds.
Rollout Expected After Fine-Tuning
The state government plans to deploy the AI-driven system more aggressively once testing and optimisation are complete. Authorities believe that integrating AI into GST enforcement will significantly enhance the department’s ability to prevent revenue leakage and dismantle organised fraud networks.
This initiative reflects a broader shift towards technology-led tax administration, aimed at improving compliance, transparency, and efficiency in the GST regime.
Read More: Pre-Deposit Field Now Made editable At The Time Of Filing Appeal: GSTNÂ
