As countries push forward with smart healthcare initiatives, they face challenges such as potential biases, lack of transparency in AI software used in clinical applications, and issues related to data security and privacy. On October 5th, the Ministry of Health and Welfare (MOHW) announced the establishment of three new AI centers: the "Responsible AI Execution Center," the "Clinical AI Certification and Verification Center," and the "AI Impact Research Center." These centers aim to address the key challenges of AI application in clinical settings—implementation, certification, and reimbursement—through cross-hospital collaboration across multiple levels and systems. This marks a significant milestone in Taiwan's AI healthcare application, representing a forward-looking policy for advancing smart healthcare.
Overview of the Three AI Centers' Support Plan
The three AI center support programs were announced on July 23 and included nationwide educational training, with submissions closing on August 1. There was strong participation from hospitals at all levels, with 30 hospitals submitting a total of 48 projects. After a three-stage evaluation by local and international experts, 19 projects from 16 hospitals were selected, including key hospitals such as National Taiwan University Hospital (NTUH), Taichung Veterans General Hospital (TVGH), Taipei Veterans General Hospital (VGH), National Cheng Kung University Hospital (NCKUH), China Medical University Hospital (CMUH), Tri-Service General Hospital (TSGH), and Linkou Chang Gung Memorial Hospital.
Minister Tai-Yuan Chiu: The Three AI Centers Aim to Improve Healthcare
Minister of Health and Welfare, Tai-Yuan Chiu, stated that the primary purpose of establishing the three AI centers is to address the issues of smart healthcare implementation, certification, and reimbursement, bringing advanced and reliable medical services to the Taiwanese people. At the same time, the centers will optimize long-term care policies, improve the healthcare environment, and strengthen resource coordination, ensuring the sustainable development of national health insurance (NHI) and smart healthcare. The centers will also help address the shortage of healthcare professionals, accelerate the integration of new drugs and technologies, and maintain Taiwan’s competitive edge in the global healthcare arena.
Director Chien-Chang Lee: Overcoming Challenges in Certification and Reimbursement
Director of the Information Technology Department, Chien-Chang Lee, highlighted that AI products in major hospitals often take more than a year to gain certification. Given the rapid pace of AI advancements, this certification process is seen as too slow. Moreover, assessing how key AI applications can secure reimbursement through scientific evaluation and verifying clinical effectiveness are major challenges in promoting smart healthcare.
To address these issues, advanced countries have introduced regulatory frameworks and consensus guidelines, such as the World Health Organization's six principles for responsible AI, the EU AI Act, and the U.S. rules for algorithmic transparency and information sharing. However, these policies have yet to be fully implemented as core drivers of progress in many countries.
The Three AI Centers: Key Solutions for Smart Healthcare's Final Hurdle
In anticipation of future needs, the MOHW has guided Taiwanese hospitals in establishing three types of AI centers to address the final hurdles in smart healthcare deployment:
Responsible AI Execution Center: This center focuses on developing management measures that align with data security and privacy protection, publicly disclosing key information such as AI models, data, and performance, and providing interpretable analytical results to enhance AI transparency. It also establishes AI model lifecycle management to ensure the continued reliability of AI in clinical use.
Clinical AI Certification and Verification Center: In collaboration with the Food and Drug Administration (FDA), this center tackles the lengthy certification process for medical AI software in Taiwan and the difficulty of gathering representative datasets for verification. Previously, external AI validation mechanisms lacked cross-hospital collaboration and dedicated institutions or personnel. This center will establish a hospital alliance with permanent institutions and specialists to integrate cross-system and cross-level electronic medical records, assisting manufacturers in validating AI models more quickly and accurately for the Taiwanese population. This will accelerate product commercialization and allow the public to benefit from high-quality smart healthcare products faster.
AI Impact Research Center: In collaboration with the National Health Insurance Administration, this center will address the challenges of securing NHI reimbursement for AI medical applications. Due to the difficulty in evaluating AI product pricing through traditional manufacturing cost assessments, the center will establish cross-system, cross-level clinical trial mechanisms. Multi-disciplinary teams will design research to evaluate the clinical benefits of AI, ensuring that AI products demonstrate both clinical effectiveness and health benefits, while also creating a scientific basis for pricing.
The launch of these three AI centers signifies Taiwan’s significant progress and forward-thinking development in the field of smart healthcare. The MOHW will continue to drive the development of AI technology, laying a solid foundation for the thriving growth of Taiwan’s smart healthcare industry.
Resource (mandarin): 智慧醫療進展速度快 衛福部成立三大AI中心推前瞻政策
