How Hospitals Can Use AI Without Exposing Patient Data

Can Use AI in Healthcare Without Putting Patient Privacy at Risk, Taiwan’s NSYSU Research Finds

6-Feb-2026 4:15 AM EST, by iGroup

Newswise — Artificial intelligence is increasingly used in healthcare to support diagnosis, disease prevention, and medical research. Yet despite its promise, many hospitals remain cautious. Patient data is among the most sensitive information any institution handles, and concerns over privacy, regulation, and data misuse often limit how medical data can be shared or analyzed, especially when cloud technologies are involved. 

A research team at National Sun Yat-Sen University (NSYSU), led by Chun-I Fan, Distinguished Professor in the Department of Computer Science and Engineering, has developed a privacy-preserving medical data warehouse system designed to address these concerns. Instead of centralizing raw patient records or granting broad cloud access, the system enables healthcare institutions to securely manage, share, and analyze medical data while protecting sensitive information. 

What sets this work apart is its focus on real-world usability. While privacy-preserving AI concepts such as encryption and federated learning are well established in theory, integrating them into existing hospital systems is far more complex in practice. NSYSU’s approach brings together international healthcare data standards, including FHIR, with cryptographic protection mechanisms and federated learning. This approach enables hospitals to collaborate and jointly train AI models, without requiring raw patient data to leave each institution’s control. 

In practical terms, AI models can “learn” from data across multiple hospitals while patient records remain encrypted and locally governed, helping institutions comply with regulatory requirements and reduce privacy risks. The system is also designed to be modular and cloud-agnostic, offering flexibility for hospitals and regional healthcare providers that face cost constraints, vendor lock-in concerns, or strict compliance obligations. 

Beyond technical design, the research emphasizes implementation experience. Through prototype development and pilot-oriented testing, the team addressed challenges such as interoperability across different systems, encrypted data operations, access control, and cross-institution collaboration—bridging the gap between academic research and operational healthcare environments. 

As healthcare systems worldwide search for responsible ways to adopt artificial intelligence, this work demonstrates that advancing medical innovation does not require compromising patient privacy. Readers are invited to find out more by exploring the full press release and watching the accompanying research video. 

Original Link: https://www.newswise.com/articles/how-hospitals-can-use-ai-without-exposing-patient-data