MUFG Bank Integrates Private AI’s De-Identification Technology to Bolster ‘OCEAN’ Big Data Platform
MUFG Bank has formally adopted a data de-identification solution provided by Toronto-based Private AI to enhance its proprietary big data infrastructure, "OCEAN." The strategic move is designed to accelerate the bank’s utilization of unstructured data while maintaining rigorous privacy governance.
Unlocking Unstructured Data for Generative AI
MUFG Bank has integrated Private AI’s solution to securely process unstructured data assets—such as emails, call center voice logs, internal documents (PDFs), and chat histories—which have traditionally been difficult to leverage due to privacy concerns.
By automatically detecting and redacting Personally Identifiable Information (PII) and sensitive data with high precision before it enters the "OCEAN" data lake, MUFG aims to enable safe, cross-sectional data analysis. Crucially, this infrastructure paves the way for the secure deployment of Generative AI and operational AI models using actual business data without compromising compliance.
Strategic Focus on Data Sovereignty and Security
Addressing the dual challenges of advancing AI adoption and ensuring strict data governance, the solution allows MUFG to de-identify data in real-time within an on-premise or closed environment. This architecture ensures that sensitive data is not transmitted to external cloud providers, mitigating third-party risk.
Validation and Future Deployment
Prior to full-scale adoption, MUFG conducted extensive technical verification (PoC) using diverse unstructured files mirroring actual operations. The bank confirmed the system's ability to accurately anonymize critical financial data categories—including names, addresses, phone numbers, bank account numbers, and insurance ID numbers—at a level sufficient for enterprise deployment.
Following this integration, MUFG plans to expand the use of the "OCEAN" platform to multiple business areas. Key initiatives include:
- Operational Enhancement: Upgrading call center workflows.
- Risk Management: Advancing fraud detection and risk assessment capabilities.
- Internal Efficiency: Promoting the use of internal knowledge bases.
About Private AI
Founded in 2019 by experts in privacy and machine learning, Private AI is a technology firm headquartered in Toronto, Canada. Its proprietary algorithms are capable of detecting and redacting over 50 types of PII across more than 52 languages. Amidst rapidly increasing corporate data flows, Private AI facilitates compliance with regulations such as the GDPR and Japan’s APPI (Act on the Protection of Personal Information). The company entered the Japanese market in September 2023 and was selected as a Gartner "Cool Vendor" in the privacy sector.
Norbert Gehrke
