Exploring Privacy Computing Technologies in the Open Financial Ecosystem

Introduction Privacy computing enables secure data value circulation while protecting privacy, gaining momentum due to technological advances and supportive policies. Financial institutions can combine external data (e.g., telecom, government, credit) with privacy‑preserving techniques to offer trustworthy aggregated services.

1. Opportunities and Challenges in Open Financial Ecosystem The open finance model requires richer external data, leading to three shifts: broader user scope, expansion from banking services to ecosystem scenarios, and complex value‑exchange environments. These changes raise privacy‑security risks that demand stronger technical safeguards.

2. Overview of Privacy Computing Privacy computing encompasses many techniques—trusted execution environments (TEE), multi‑party computation (MPC), federated learning, homomorphic encryption, differential privacy, and zero‑knowledge proofs. Each has distinct performance, security, and applicability characteristics, and no single method satisfies all needs.

3. Integration in the Full Data‑Analysis Process The workflow includes:

Requirement exploration: understanding business needs and data availability, assessing whether external data is required.

Data processing: aggregating data from multiple providers, performing alignment, and applying privacy‑preserving intersection methods such as hash‑based PSI, obfuscation circuits, fully homomorphic encryption, or oblivious transfer.

Joint modeling/computation: using federated learning (horizontal, vertical, transfer) and MPC for secure model training and inference.

Model evaluation/optimization: leveraging MPC, TEE, and AI techniques to improve accuracy while preserving privacy.

Blockchain integration: recording protocol steps, results, and incentives on-chain to ensure auditability and trust.

4. Financial‑Sector Use Cases

User asset‑level view: cross‑institution asset calculation using MPC and blockchain for provenance.

Joint marketing scoring: multi‑party data orchestration with on‑chain evidence of process compliance.

Joint risk control: linking corporate, tax, social‑security, and utility data across banks and government agencies via privacy computing to improve credit risk assessment.

Contract comparison with TEE‑based SaaS: encrypting contracts on the client side, processing them inside a trusted enclave, and returning encrypted results, all logged on blockchain.

5. Future Outlook The field will continue to evolve through parallel development of multiple techniques, algorithmic optimization, hardware acceleration, standardization efforts, and increased interoperability. Integration with cloud platforms, AI models, and IoT edge devices will further expand the applicability of privacy computing.

6. Q&A The speaker addressed data‑authorization via blockchain, highlighted the most used technologies in their bank (TEE, MPC, federated learning), and discussed ongoing research into homomorphic encryption, differential privacy, and zero‑knowledge proofs.