MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202621049804 A) filed by Rahul Chelani; Supriya Shrivastava; Nidhi Tiwari; Apoorv Khare; Ashish Shrivas; and Pranija Kasture on April 19, 2026, for A Privacy-Preserving Artificial Intelligence System For Cross-Institution Money Laundering Detection Using Encrypted Graph Neural Networks And Zero-Knowledge Proofs.
Inventors include Rahul Chelani; Supriya Shrivastava; Nidhi Tiwari; Apoorv Khare; Ashish Shrivas; and Pranija Kasture.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: A privacy-preserving system and method are disclosed for detecting money laundering across multiple financial institutions. The system includes a behavioral embedding module configured to extract transaction features from financial accounts to generate behavioral embeddings. An encryption module encrypts the embeddings using a homomorphic encryption scheme that enables computation on encrypted data without decryption. A secure aggregation coordinator receives encrypted embeddings from multiple institutions and constructs a cross-institution encrypted transaction graph representing relationships among accounts while preserving confidentiality. An encrypted graph neural network processes the encrypted graph to identify anomalous financial behavior. A temporal causal chain detection module analyzes transaction timing, amounts, and behavioral similarity across institutions to identify multi-stage laundering sequences. A zero knowledge compliance verification module generates cryptographic proofs verifying compliance with anti-money laundering rules without revealing transaction-level information. The system enables secure cross-institution analysis while preserving confidentiality of financial data.
Disclaimer: Curated by HT Syndication.
