Exploring the integration of privacy-enhancing technologies in Taiwan’s artificial intelligence and data protection frameworks

Abstract

This paper explores the integration of privacy-enhancing technologies (PETs) within Taiwan’s artificial intelligence (AI) and data protection frameworks, focusing on the Taiwan AI Fundamental Act and the Personal Data Protection Act. In response to the Constitutional Court’s directive for improvements in data protection, the Ministry of Digital Affairs proposed PET guidelines to align technology with regulatory frameworks. PETs, such as differential privacy, homomorphic encryption and secure multiparty computation, are highlighted as innovative solutions that balance data privacy with usability. Despite their potential, PETs face significant technical, legal and economic challenges. Implementing these technologies is particularly complex in sensitive sectors such as finance and healthcare, where data security is paramount. Additionally, the legal ambiguity surrounding the application of PETs creates a risk of inconsistent interpretation and enforcement, especially across different industries. While some degree of ambiguity is necessary for technological flexibility, the author argues that clear legal frameworks and technical guidelines must be provided to support the practical implementation of PETs. The government must actively promote education and cross-sector collaboration to ensure the consistent application and successful adoption of these technologies. Although the PETs guideline could reduce the efforts for implementation of PETs, applying the guideline might be an issue for decision makers. Thus, while using suggested techniques, one should have a basic knowledge of these technologies in order to gain maximum value from the guideline. On the other hand, if researchers, especially legal researchers seeking legal compliance, had a deeper understanding of PETs’ meaning and the results after implementing, it might help reduce the need for data protection requirements in practice. Ultimately, the author concludes that balancing privacy protection with technological innovation is vital to ensuring the long-term success of PETs in Taiwan’s evolving data governance landscape.