Exploring Taiwan’s AI Applications in Healthcare

Taiwan is globally recognised as a leader in the semiconductor industry, renowned not only for its advanced manufacturing capabilities but also for its comprehensive integrated supply chain, from design and fabrication to packaging and testing. The rapid emergence of artificial intelligence technologies, particularly those using generative AI, has triggered an unprecedented surge in demand for advanced semiconductor chips. The cutting-edge applications made possible by AI require increasingly sophisticated and high-performance chips. At present, Taiwan’s fully developed and sophisticated local supply chain gives its semiconductor industry a distinct competitive advantage over those of other regions.

Similarly, the high standards of medical care in Taiwan have earned it significant worldwide recognition in recent years. The global quality-of-life database Numbeo has ranked Taiwan number one in the world for eight consecutive years in its Health Care Index. The Health Care Index evaluates multiple factors, including the technical skills and competencies of medical personnel, the efficiency with which medical examinations and reports are completed, the quality of medical equipment and software, and the accuracy of medical reports that collectively reflect the overall quality of healthcare services. Moreover, Taiwan’s performance during the COVID-19 pandemic reinforced Taiwan’s standing as a worldwide leader in medical standards and public health management.

Taiwan’s competitiveness in both the semiconductor and medical sectors has positioned the island as an ideal location for developing AI applications within the healthcare industry.

Taiwan’s National Health Insurance system

The medical system in Taiwan is structured by a comprehensive insurance framework known as National Health Insurance (NHI), which is a compulsory social insurance programme designed to provide universal health coverage to Taiwanese citizens.

Under the National Health Insurance Act, NHI is exclusively managed by a single governmental agency, the National Health Insurance Administration (NHIA). This centralised management structure is a key factor in the system’s overall efficiency and adaptability. According to recent data, approximately 99.9% of the Taiwanese population is enrolled in NHI. Benefiting from the NHI system, Taiwanese citizens enjoy affordable access to high-quality medical treatment.

However, demographic trends are likely to bring new challenges to the NHI system in the near future. Like many countries around the world, Taiwan is experiencing a rapidly ageing population. Societal shifts, including delayed marriage and low fertility rates, have contributed to a significant demographic transformation within Taiwan, with the island being classified as an “aged society” (in which 14% of the population is aged 65 or over) starting in 2018, and it was classified as a “super-aged society” (in which more than 20% of the population is aged 65 or over) at the end of 2025.

As the proportion of elderly individuals rises, the demand for healthcare services is expected to increase correspondingly. As such, Taiwan’s healthcare system is currently facing a critical workforce shortage, which complicates efforts to meet the escalating needs of its ageing population. In response, Taiwan is exploring a range of strategies, including integration of advanced technologies, enhancement of workforce training programmes, and implementation of policy reforms aimed at optimising resource allocation.

AI applications in healthcare development in Taiwan

In the context of healthcare development in Taiwan, AI applications have been frequently integrated into assistance and alert functionalities. These foundational uses leverage AI’s inherent capabilities to analyse complex datasets of medical data, thereby enabling tailored applications to address specific clinical needs.

Medical management, diagnosis and prediction-based diagnosis

More recently, AI’s role in healthcare has expanded to encompass medical management, diagnosis and prediction-based diagnosis. In the area of medical management, AI has been applied to optimise resource utilisation and enhance the organisation of outpatient services. In diagnostic applications, AI-powered identification systems have been employed to facilitate faster and more precise image analyses, thereby supporting clinicians in achieving more accurate diagnoses. Regarding prediction-based diagnosis, by aggregating and analysing extensive datasets including patient histories, genetic information and real-time health indicators, AI can generate risk assessments and recommend precautionary measures to individual patients.

In the realm of digital healthcare, to address the shortage of healthcare workers, the Industrial Technology Research Institute (ITRI) has launched the MedBobi 2.0 Medical Voice Assistant system, which integrates clinical language data with multimodal generative AI technology. This innovative system integrates clinical language with multimodal generative AI technology, enabling health professionals to quickly generate medical reports using voice commands alone. This system is expected to reduce administrative workloads by approximately 75%, significantly enhancing healthcare efficiency.

In addition, the MedBobi 2.0 system performs in-depth analysis based on patient records and medical images, assisting physicians in providing diagnostic recommendations and strengthening clinical decision-making. The system has been successfully implemented in clinical settings such as Far Eastern Memorial Hospital, Chung Shan Medical University Hospital, Kaohsiung Veterans General Hospital and Chi Mei Hospital, demonstrating Taiwan’s comprehensive capabilities in AI healthcare from research to practical application. Through applications such as these, AI supports clinicians in delivering more accurate diagnoses and personalised treatment plans.

Clinical care

In the area of clinical care, AI systems can aggregate and analyse vast amounts of patient biodata, including medical histories, imaging data and real-time physiological measurements. Through continuous monitoring, AI can detect subtle changes or emerging risks in a patient’s condition. Such uninterrupted surveillance facilitates timely identification of potential complications, allowing for proactive interventions that lead to better patient outcomes.

Beyond individual patient monitoring, AI-based analysis of large-scale datasets enables the detection of non-compliance or deviations from established medical standards. Through comparison monitoring, AI can identify instances where a patient may not be following prescribed instructions, or where clinical practices diverge from accepted standards. Corresponding alerts could serve as safeguards to prevent avoidable clinical errors.

A practical example of AI applications in clinical care is demonstrated by technologies developed at National Cheng Kung University Hospital, where AI-driven tools calculate muscle composition ratios based on CT scans. This innovation enables early warning of potential side effects associated with chemotherapy.

R-AI Hub

In 2024, the Ministry of Health and Welfare launched the Responsible AI Execution Center (referred to as the "R-AI Hub"), aiming at establishing actionable clinical standards for AI applications. The R-AI Hub has three core objectives:

• to establish an independent review committee responsible for reviewing proposed AI products to ensure compliance with privacy regulations, cybersecurity standards, and demonstrable clinical benefits;
• to develop transparency and explainability analyses to eliminate concerns surrounding the so-called algorithmic black box; and
• to promote localised testing and comprehensive lifecycle management.

Given the high sensitivity of AI models to data, even internationally certified models may produce errors due to differences with local data characteristics in Taiwan, such as prevalence rates or population characteristics.

Legal framework governing healthcare and AI in Taiwan

Regulations on artificial intelligence

In 2019, the Taiwan government began drafting the Artificial Intelligence Development Fundamental Act, with ongoing refinement continuing since then. On 14 January 2026, the Artificial Intelligence Basic Act officially took effect, establishing a foundational legal framework for the development, application, and governance of artificial intelligence in Taiwan. The Artificial Intelligence Basic Act explicitly defines artificial intelligence as a system capable of autonomous operation, which, through input or sensing and by means of machine learning and algorithms, can generate outputs such as predictions, content, recommendations, or decisions that affect physical or virtual environments to achieve explicit or implicit objectives. The Act provides clarity and consistency regarding its scope and application. At its core, the legislation is intended to strike a balance between encouraging AI innovation and protecting fundamental rights, such as human dignity, privacy and equal treatment.

The Artificial Intelligence Basic Act can be described as a framework of principles, rather than a prescription of detailed regulatory mandates. Within this legal framework, it is anticipated that relevant competent authorities will develop and implement sector-specific measures to operationalise the principles set forth by the Act. Given the Act’s recent enactment, some time will be required for rules to be put into practice and to observe how the Act affects Taiwan’s AI development.

Management of medical devices

Across the healthcare industry, AI applications are increasingly being integrated into various diagnostic, therapeutic and monitoring tools. In Taiwan, such AI-driven healthcare solutions meeting specific requirements will be classified as medical devices.

The Medical Devices Act provides primary regulations governing medical devices in Taiwan. Management of AI applications in the context of healthcare is likely to fall under the purview of the Medical Devices Act. According to the Medical Devices Act, the term “medical device” refers to instruments, machines, apparatuses, materials, software, reagents for in vitro use and related articles thereof, whose design and use achieve one of the following primary intended actions in or on the human body by means other than pharmacological, immunological, metabolic or chemical means:

• diagnosis, treatment, alleviation or direct prevention of human diseases;
• modification or improvement of the structure and function of the human body; and
• control of conception.

The Ministry of Health and Welfare (MOHW) is the competent authority overseeing the matters related to healthcare, products and industries. Under the MOHW, the Food and Drug Administration (TFDA) supervises the regulation of food, drugs, medical devices, and cosmetics to ensure their safety and quality. The TFDA is responsible for granting product registrations and approving clinical trials, as well as monitoring manufacturing processes and imports. From a legal standpoint in Taiwan, the manufacture or import of medical devices is permitted only after the MOHW has issued a medical device permit licence which grants registration and market approval.

Further, to accommodate emerging and advanced categories of medical devices, the TFDA issued the Guidance for the Inspection and Registration of Medical Software of AI/ML-based Technologies (the “Guidance”) in August 2021. The Guidance describes the inspection and registration requirements for medical software and medical devices that use AI/ML-based (artificial intelligence/machine learning-based) technologies. Based on the list of AI/ML-based medical devices approved as of February 2026, approximately 89 domestically manufactured AI/ML-based medical devices and approximately 148 imported AI/ML-based medical devices had been authorised by the TFDA for manufacture or import into Taiwan. Such AI/ML-based medical devices include ultrasound devices with AI-assisted interpretation, complementary diagnosis software and X-ray photography systems.

It is foreseeable that the integration of AI into healthcare, including diagnostics and treatment, will continue to expand significantly. These advancements promise to enhance the accuracy and personalisation of healthcare delivery. However, the proliferation of AI applications in this sector also introduces complexities in regulatory oversight and will present additional challenges at both national and international levels.

Personal data protection

In addition to the management of medical devices, the protection of personal data has emerged as a key concern in the deployment of AI within the healthcare field. As AI technologies rely on vast amounts of sensitive patient information to enhance diagnostic accuracy and treatment personalisation, important issues regarding the protection of personal data have arisen.

The Personal Data Protection Act

The Personal Data Protection Act (PDPA), last amended in November 2025 and set to take effect at a date yet to be announced, serves as the core legislation for personal data protection in Taiwan. Under the PDPA, “personal data” encompasses any information capable of directly or indirectly identifying an individual, such as name, date of birth, national identification card number, passport number, physical characteristics, fingerprints, marital status, family information, educational background, occupation, medical records, healthcare data, genetic data, sexual history, records of physical examination, criminal record, contact information, financial condition and social activities.

The PDPA outlines the responsibilities of data controllers and processors in the collection, processing, and use of personal data, ensuring that such activities are conducted in accordance with the regulations, such as acquiring the consent of the data subject prior to the use and collection of personal data. The PDPA also prescribes requirements for personal data security, mandates timely notification in the event of data breaches, and provides mechanisms for individuals to exercise their rights regarding their personal data.

Additionally, the scope of personal data is further delineated to include a distinct category of “sensitive personal data”, which encompasses information such as medical records, healthcare details, genetic data, sexual history, physical examination results and criminal record. Due to the nature of such data, the PDPA imposes more stringent regulatory requirements, mandating that such sensitive personal data may be collected, processed, and used only under specific conditions, such as where the collection, processing, and/or use is/are expressly required by law, or after provision of the data subjects’ written consent.

In the context of the healthcare sector, the personal data processed by AI largely encompass medical records and healthcare-related information, which are classified as sensitive personal data under the PDPA and are subject to more stringent regulatory requirements. Moreover, the deployment of AI within healthcare frequently entails cross-border transfer of personal data, thereby introducing additional compliance requirements in data protection laws. Consequently, the collection, processing, and use of personal data in AI-driven healthcare applications will be subject to multiple layers of legal considerations and require more diligence in implementing data compliance mechanisms to mitigate risks and protect personal data.

Initiatives regarding personal health data

Since its inception in 1995, Taiwan’s NHI system has been managed by the NHIA, which oversees a large volume of personal data. From 2000 to 2016, the NHIA entrusted data management to the National Health Research Institute, which created the National Health Insurance Research Database for external research. In 2012, amid rising concerns about personal data privacy, seven individuals objected to the NHIA’s disclosure of their personal data to third parties, resulting in petitions and administrative lawsuits that were ultimately unsuccessful. In 2017, the plaintiffs pursued a constitutional interpretation to determine the legality of such personal data release.

In August 2022, Taiwan’s Constitutional Court ruled in the Judgment of Constitutional Court (111) Sian-Pan-Zi No 13, mandating that laws must be revised within three years to strengthen protection of personal data under the PDPA. Key requirements included establishing an independent oversight mechanism, defining clear regulations for NHI data usage, and permitting individuals to opt out of personal data usage.

To comply, the PDPA was amended in May 2023, designating the Personal Data Protection Commission (PDPC) as the authority overseeing the relevant data protection regulations. A Preparatory Office for the PDPC was accordingly established in December 2023 for the purpose of formulating organisational laws and regulations of the PDPC and revising and interpreting the PDPA. However, despite the expiration of the time constraint imposed by the Constitutional Court, the PDPC has not yet been officially established, as the organisation act governing its formation is still under review.

Meanwhile, the PDPA was further amended in November 2025. The 2025 amendment primarily focuses on revising the wording of the PDPA provisions in light of the PDPC’s anticipated powers and responsibilities upon its forthcoming establishment. Additionally, the amended PDPA enhances oversight of public agencies by requiring them to appoint Data Protection Officers, and introduces provisions for the PDPC to co-ordinate and co-operate with competent authorities from the central government as well as local governments in supervising non-public entities. Given that the PDPC has yet to be formally established, the enactment of the amended PDPA remains pending further official announcement.

Additionally, in response to the issues raised by the Judgment of the Constitutional Court (111) Sian-Pan-Zi No. 13, a draft legislation, titled the National Health Insurance Data Management Act, was published by the Executive Yuan in May 2025, introducing comprehensive measures to enhance the governance of health insurance data. The draft Act would mandate the establishment of an advisory council tasked with providing expert counsel and overseeing the implementation of related policies. It further defines a framework for the construction, management, and evaluation of pertinent data repositories, as well as standardised procedures for transmission of data by insurers to competent authorities. Moreover, the draft Act would also entitle data subjects to opt out of data usage.

Following review, the National Health Insurance Data Management Act was passed by the Legislative Yuan on December 2, 2025. The date of enactment has yet to be announced by the Executive Yuan. Such legislation is expected to ensure that the use of health insurance data beyond its originally intended purposes is handled securely and reasonably in accordance with legal procedures. This will maximise the value of health insurance data applications while safeguarding individuals’ rights to information privacy and autonomy.

Conclusion

Given that healthcare inherently involves human dimensions and places significant emphasis on ethical considerations, it is imperative to differentiate the application of AI within this sector from its use in other fields. As AI technologies continue to advance, the healthcare sector faces increasingly complex legal and ethical challenges. Among these, the lack of transparency in AI decision-making processes raises concerns regarding accountability and the attribution of responsibility. Consequently, existing legal frameworks may be inadequate to address the issues arising from AI integration in healthcare. In light of the above, the development of regulatory frameworks governing AI deployment in healthcare must proceed with diligence and specificity, ensuring that considerations of patient safety, personal privacy, and ethical standards remain paramount.