In recent findings from a parliamentary committee, it has been revealed that the finance ministry has successfully integrated artificial intelligence (AI) across various systems designed to combat cybercrime and financial fraud. This integration is particularly evident within tax and financial intelligence networks, where AI tools facilitate real-time monitoring, anomaly detection, and risk scoring. However, it is noteworthy that the Directorate of Enforcement (ED), the agency responsible for investigating money laundering, has yet to implement AI technologies.
AI Applications in Financial Intelligence
The Financial Intelligence Unit-India (FIU-IND) employs a variety of AI-driven strategies, including:
- FINnet 2.0: This system leverages AI and machine learning (ML) techniques to generate entity linkages, create enriched profiles, and visualize networks, thereby unveiling concealed relationships.
- Entity Resolution Tools: These tools extract and match entities across datasets and enhance data using external databases, which helps in identifying potential relationships between entities.
- Law Enforcement Agency (LEA) Prediction Models: These utilize pattern recognition and classification techniques, based on historical data, to forecast trends.
- Transaction Monitoring Systems: These systems flag suspicious patterns, including structured transactions, the use of shell companies, and transactions with unusual values.
- Dynamic Risk Assessment Systems: AI can be integrated into these systems to adjust thresholds in real time, based on behavioral patterns, risk profiles, and trends in fraud.
AI’s Role in Tax Fraud Detection
- Risk Assessment Systems: AI and ML are employed to flag fraudulent Input Tax Credit (ITC) claims and uncover networks involved in tax evasion.
- Data Analytics Systems: Utilizing third-party data, these systems aid investigations and enhance risk scoring.
- Compliance Monitoring Dashboards: AI can power these dashboards to assign real-time risk scores to businesses, facilitating proactive inspections.
- Fraud Detection Systems: These systems are vital for identifying fake invoices, circular trading practices, and suspicious transaction patterns.
- Registration Risk Scoring Systems: High-risk applications are routed for biometric Aadhaar authentication and field verification.
- Centralized Analytics Portal: This portal processes AI outputs seamlessly at the backend, allowing officers to utilize simplified dashboards for easier access.
- Proactive Detection Capabilities: AI systems hold the promise of identifying fake invoices and fraudulent refund claims before any revenue is compromised.
Lack of AI in ED Investigations
The Directorate of Enforcement has confirmed to the committee that, currently, the ED does not deploy any AI tools for its money laundering investigations. The agency relies on conventional methods and has not yet developed or implemented AI-based models. Nonetheless, it is in the process of monitoring trends in AI-enabled fraud and is collaborating with FIU-IND and other agencies to investigate shell companies and suspicious transactions.
Moreover, the ED has indicated potential applications for AI, such as analyzing large datasets, prioritizing high-risk cases, and expediting investigations. The agency acknowledged that AI could link corporate records, banking data, land records, and tax information to unearth hidden income. However, any future deployment would necessitate evidentiary compliance and must be conducted under human supervision.
Conclusion
The integration of AI in financial intelligence showcases promising advancements in detecting and preventing financial crimes. While the finance ministry actively employs AI tools, the absence of such technologies in the Directorate of Enforcement remains a gap that could be addressed in the future. As AI continues to evolve, its potential in improving the efficiency and effectiveness of financial investigations will likely gain greater recognition.
Also read
