Taiwan - Kaohsiung - Smart Lighthouse AI Program

Overview

The Kaohsiung Smart Lighthouse AI Program, also referred to as the Kaohsiung Intelligent Lighthouse Program or Smart Kaohsiung Lighthouse Project, represents a significant undertaking by the Kaohsiung City Government to embed artificial intelligence into its urban governance framework. This initiative is a core component of the broader 'Kaohsiung Sovereign AI Project,' which was officially launched to enhance urban governance efficiency, strengthen smart city development, and foster new business opportunities through advanced AI applications. The program focuses on developing Taiwan's first sovereign AI platform, envisioned as a 'city governance brain' powered by physics-informed AI, to address critical urban challenges such as traffic management, facility anomaly detection, and disaster management. It is designed to be tailored to Kaohsiung's specific needs, offering civil servants an AI assistant capable of analyzing citywide data and recommending rapid, data-driven responses.

While the Kaohsiung Smart Lighthouse AI Program itself is an operational initiative rather than a standalone legislative act, its implementation is deeply intertwined with and governed by Taiwan's evolving national regulatory framework for artificial intelligence and data protection. Key among these is the Artificial Intelligence Basic Act, which was passed by Taiwan's Legislature in December 2025, providing a comprehensive legal framework for the development, application, and regulation of AI technologies across the nation. Additionally, the program operates under the strictures of Taiwan's Personal Data Protection Act (PDPA), ensuring the safeguarding of personal information collected and processed by AI systems. The program also aligns with Kaohsiung City's commitment to sustainability, complementing the city's Net Zero Development Ordinance and related decarbonization plans.

The program emphasizes strategic partnerships with industry leaders like NVIDIA and Chunghwa Telecom, as well as academic institutions, to ensure the autonomy, reduce biases, and enhance the capabilities of its generative AI platform. It aims to position Kaohsiung as a model smart city, demonstrating leadership in AI-driven sustainability, digital twins, and intelligent municipal governance, with the potential for its models to be expanded to other cities or exported internationally.

Definitions

While the Kaohsiung Smart Lighthouse AI Program itself does not introduce new legal definitions, its operations are guided by definitions established in national legislation, particularly the Artificial Intelligence Basic Act and the Personal Data Protection Act (PDPA). Key terms relevant to the program include:

• Artificial Intelligence (AI): The Artificial Intelligence Basic Act, as it passed, is expected to provide a broad definition covering various AI techniques and approaches, from knowledge-based algorithms to sophisticated neural networks, ensuring comprehensive regulatory coverage. This definition would encompass the generative AI platform and physics-informed AI systems utilized by the Kaohsiung Smart Lighthouse AI Program.
• Personal Data: Under Taiwan's Personal Data Protection Act (PDPA), 'personal data' is broadly defined to include name, date of birth, I.D. card number, passport number, characteristics, fingerprints, marital status, family information, education, occupation, medical records, genetic information, and sexual-life information. The collection, processing, and use of such data are generally subject to notice and consent requirements. The AI program must adhere to these definitions when handling citizen data for urban services.
• High-Risk AI: The Artificial Intelligence Basic Act proposes a risk-based framework for AI, likely referencing international standards. This framework would classify AI technologies based on risk levels (e.g., unacceptable, high, limited risk), with varying degrees of regulatory oversight. AI applications within the Kaohsiung Smart Lighthouse AI Program that impact critical areas like public safety, traffic control, or disaster management could be categorized as high-risk.
• Data Governance: Refers to the overall management of the availability, usability, integrity, and security of data used in AI systems. The AI Basic Act and the PDPA emphasize principles of data governance, including avoiding unnecessary data collection and promoting data protection by design.
• Smart City: Although not explicitly defined in national AI legislation, the concept of a 'smart city' in Taiwan, as embodied by Kaohsiung, involves utilizing 5G AIoT applications for digital governance and digital transformation in public and private sectors to improve living environments and industrial transformation.

Governance and Institutional Framework

The Kaohsiung Smart Lighthouse AI Program is primarily driven by the Kaohsiung City Government. Since 2019, the city government has been actively formulating a blueprint for smart city governance, establishing the Kaohsiung Smart City Promotion Committee, which includes representatives from industry, government, academia, and research institutes. The Kaohsiung Smart City Project Management Office (KPMO) was also established to oversee such initiatives.

At the national level, the Artificial Intelligence Basic Act designates the National Science and Technology Council (NSTC) as the central competent authority responsible for guiding the development and use of AI. This means the NSTC will oversee the implementation of the AI Basic Act and ensure that local initiatives like Kaohsiung's AI program comply with national principles and standards. The Act also mandates the Cabinet to establish a committee, convened by the premier and comprising experts, industry representatives, agency heads, and local government leaders, to formulate a national AI development framework. This national committee would likely provide oversight and coordination for major local AI initiatives.

Regarding data protection, the Personal Data Protection Act (PDPA) is enforced by relevant data protection authorities. The draft amendments to the PDPA anticipate the establishment of a Personal Data Protection Commission, which will play a crucial role in bolstering personal data protection in the digital era. This commission would be responsible for ensuring that the data handling practices within the Kaohsiung Smart Lighthouse AI Program adhere to privacy regulations.

Furthermore, the Ministry of Transportation and Communications (MOTC) has established a Transportation AI Promotion Committee, which convened its first meeting in August 2025. This committee aims to accelerate the integration of AI into Taiwan's transportation systems and is drafting internal AI implementation regulations in line with the Executive Yuan's AI Basic Act. Given the Kaohsiung program's focus on traffic management, the MOTC's guidelines and frameworks will be highly relevant.

Key Focus Areas

The Kaohsiung Smart Lighthouse AI Program, as a cornerstone of Kaohsiung's smart city vision, has several key focus areas:

• AI-driven Urban Governance: The program aims to develop a sovereign AI system to support various government operations. This 'city governance brain' will leverage advanced computing power and AI models to enhance efficiency and urban resilience. Applications include traffic management, facility anomaly detection, and disaster management.
• Sustainability and Net-Zero Initiatives: A strong emphasis is placed on sustainability and achieving net-zero emissions. The program complements Kaohsiung's Net Zero Development Ordinance and integrates AI into real-world carbon neutrality projects. This includes developing smart net-zero technologies and policies.
• Digital Innovation and Integration: The program fosters deep AI & ICT Integration through collaborations with industry giants like NVIDIA and Chunghwa Telecom. It involves the ambitious Kaohsiung AI Upgrade 2.0 master project, targeting infrastructure and sustainability, and focuses on Data Security & Cloud Transformation for intelligent municipal governance.
• Public Safety and Disaster Prevention: AI applications are utilized for large-scale event planning, traffic flow analysis, and disaster prevention, providing customized response strategies. This includes predicting floods and optimizing emergency responses. Earlier projects in Kaohsiung also utilized AI for smart signals to optimize traffic flow and 3D public safety inspection and warning systems using AI image analysis and sound recognition to detect potential criminal activity.
• Smart Transportation: A key aspect involves optimizing traffic signals and managing traffic flow using AI. This aligns with broader national efforts by the Ministry of Transportation and Communications to increase AI applications in transportation management, focusing on improving service efficiency and traffic safety.
• Data Openness and Governance: The program operates within a framework that promotes data autonomy and localized AI models. It is expected to align with national efforts to establish mechanisms for open data, data sharing, and data reuse to improve the availability of high-quality, non-sensitive data for AI models while protecting personal data.

Implementation Framework

The implementation of the Kaohsiung Smart Lighthouse AI Program is multi-faceted, involving technological development, strategic partnerships, and integration with existing urban infrastructure. At its core is the development of a sovereign AI platform, which is Taiwan's first, designed as a 'city governance brain' to process and analyze vast amounts of urban data. This platform integrates data from infrastructure and disaster-prevention systems, creating a constantly updated virtual model of the city.

Strategic partnerships are crucial for its implementation. Collaborations with global and domestic enterprises such as NVIDIA and Chunghwa Telecom, alongside state-owned enterprises, are aimed at ensuring the autonomy of the AI models and reducing biases, while enhancing computing power and AI capabilities. The program also benefits from strengthened Academia & Innovation partnerships, fostering a talent pool and research in AI.

The program is being rolled out with a focus on practical applications. For instance, the AI platform will be utilized for real-world scenarios like large-scale event planning, traffic flow analysis, and industrial safety monitoring. It is also integrated into Kaohsiung's net-zero initiatives, leveraging technology to achieve carbon neutrality goals. The city's Asia New Bay Area serves as Taiwan's largest 5G AIoT demonstration zone, providing a real-world testing ground for next-generation technologies that feed into the Smart Lighthouse Program.

The implementation framework also considers the broader national AI strategy. The Ministry of Transportation and Communications (MOTC) is developing a systematic AI roadmap and governance framework for transportation, which will influence the Kaohsiung program's smart mobility aspects. The Executive Yuan's AI Basic Act, passed in December 2025, lays the groundwork for interagency collaboration and unified AI regulation, ensuring that local initiatives align with national principles.

Monitoring and Evaluation

Monitoring and evaluation of the Kaohsiung Smart Lighthouse AI Program are implicitly guided by the principles and requirements set forth in Taiwan's national AI and data protection legislation. Although specific monitoring mechanisms for this particular city program are not detailed in the provided sources, the overarching Artificial Intelligence Basic Act emphasizes accountability, transparency, and safety.

Under the AI Basic Act, the National Science and Technology Council (NSTC) serves as the central competent authority, responsible for overseeing AI development and application. This includes ensuring that AI initiatives, such as Kaohsiung's, adhere to the established guiding principles, which include privacy protection, data governance, cybersecurity, safety, transparency, explainability, fairness, and non-discrimination. The Act's proposed risk-based framework suggests that AI applications within the Smart Lighthouse Program would be subject to varying degrees of regulatory oversight and compliance burdens based on their assessed risk levels.

For data protection aspects, compliance with the Personal Data Protection Act (PDPA) is paramount. This involves ensuring that the collection, processing, and use of personal data within the AI systems avoid unnecessary data practices and promote 'data protection by design and by default.' The soon-to-be-established Personal Data Protection Commission is expected to play a key role in monitoring adherence to these data privacy standards.

The Kaohsiung City Government's own Smart City Promotion Committee and Project Management Office (KPMO) would likely conduct internal monitoring of the program's performance against its stated goals, such as improving urban governance efficiency, reducing carbon emissions, and enhancing public safety. Regular reviews and adjustments of relevant laws and regulations for data openness, sharing, and reuse are also mandated to improve data availability for AI usage while maintaining privacy.

Penalties, Liability, and Appeals

As the Kaohsiung Smart Lighthouse AI Program is an operational initiative, any penalties, liability, and appeals would fall under the broader legal framework governing AI and data protection in Taiwan, primarily the Artificial Intelligence Basic Act and the Personal Data Protection Act (PDPA).

The Artificial Intelligence Basic Act, passed in December 2025, sets a legal framework to prevent potential harm from AI applications. It requires authorities to avoid AI uses that infringe on fundamental rights (life, bodily integrity, freedom, property), undermine social order, national security, or the environment. It also guards against biased or discriminatory outcomes, false advertising, misinformation, and fabricated content. High-risk AI products or systems must carry clear notices or warnings. While the Act itself establishes the framework, specific enforcement measures, penalties, and liability provisions are expected to be detailed in subsequent regulations or through cross-ministerial coordination.

Under the Personal Data Protection Act (PDPA), non-government agencies (which would include entities involved in the program if not directly government-operated, or even government agencies in certain contexts) are liable for damages arising from any injury caused by unlawful collection, processing, or use of personal data, or other infringements on data subjects' rights due to violations of the PDPA, unless they can prove the injury was not caused by their willful act or negligence. Data subjects also have the right to complain to the relevant data protection authorities and seek compensation. The PDPA also includes administrative fines for various violations, with specific amounts depending on the nature and severity of the infringement.

For AI applications causing harm, the AI Basic Act's emphasis on accountability suggests that mechanisms for redress, potentially including civil liability for damages, would be established. The creation of a national AI development framework committee by the Cabinet, composed of experts and relevant agency heads, would also likely address issues of liability and appeals within the AI ecosystem.

Relationship to Other Instruments

The Kaohsiung Smart Lighthouse AI Program operates within a comprehensive legal and policy ecosystem in Taiwan, and its relationship to other instruments is crucial for its legitimacy and effectiveness:

• Artificial Intelligence Basic Act: This is the primary overarching legislation. Passed in December 2025, it establishes the fundamental legal framework and guiding principles for AI development and application in Taiwan. The Kaohsiung program must align with its principles of sustainability, human autonomy, privacy protection, data governance, cybersecurity, safety, transparency, fairness, and accountability. The Act designates the National Science and Technology Council (NSTC) as the central competent authority, providing national-level oversight for local AI initiatives.
• Personal Data Protection Act (PDPA): The PDPA is critical for the program's data handling practices. Any collection, processing, or use of personal data by the AI systems within the Lighthouse Program must comply with the PDPA's requirements for notice, consent, purpose limitation, data minimization, and proportionality. The AI Basic Act further advocates for integrating data protection by design and by default in AI development.
• Regulations on the Prevention and Control of Fraud and Crime Harm: Enacted in July 2024, these regulations, which require online advertising platforms to disclose the use of AI-generated personal images, demonstrate Taiwan's proactive approach to AI-related risks. While not directly governing the Lighthouse Program, they reflect a broader regulatory environment that influences how AI-generated content might be handled or perceived in public-facing smart city applications.
• Act for the Promotion of Data Innovation and Utilisation (Draft): This draft legislation, being prepared by the Ministry of Digital Affairs (MODA), aims to increase open data accessibility, provide shared data, and establish cross-industry data-sharing mechanisms. The Kaohsiung program's focus on data autonomy and localized AI models would greatly benefit from and contribute to such a national data governance framework.
• Kaohsiung's Net Zero Development Ordinance: This local ordinance, Taiwan's first, provides the framework for Kaohsiung City's decarbonization plans. The Smart Lighthouse AI Program's emphasis on sustainability and smart net-zero initiatives directly supports the goals of this ordinance, using AI to achieve greenhouse gas emission reductions.
• AI Technology R&D Guidelines (2019): Issued by the Ministry of Science and Technology, these guidelines demonstrate the government's commitment to improving Taiwan's AI R&D environment and building an AI-embedded society with human-centered values. The Kaohsiung program's emphasis on human autonomy and societal well-being aligns with these foundational guidelines.

International Alignment

Taiwan's approach to AI regulation, as exemplified by the Artificial Intelligence Basic Act and its influence on programs like the Kaohsiung Smart Lighthouse AI Program, demonstrates a clear alignment with international standards and frameworks. The draft AI Basic Act, and subsequently the enacted version, explicitly emphasizes principles that resonate with those established by international organizations and leading jurisdictions.

Key areas of international alignment include:

• OECD AI Principles: Taiwan's AI Basic Act incorporates principles such as sustainable development, human autonomy, privacy protection and data governance, cybersecurity and safety, transparency and interpretability, fairness and non-discrimination, and accountability. These principles are consistent with the OECD's recommendations on AI.
• EU AI Act: The Taiwanese framework, particularly the proposed risk-based approach for AI providers and developers, references international standards and suggests an approach similar to the EU AI Act. This involves classifying AI technologies based on risk levels (unacceptable, high, and limited risk) and applying varying degrees of regulatory oversight. The requirement for online advertising platforms to disclose AI-generated personal images also resembles EU AI Act requirements on AI-generated content and deepfake technology.
• G7 AI Principles: The guiding principles of Taiwan's AI Basic Act, including those related to human rights, privacy, industrial competitiveness, and public interest, reflect broader international consensus found in frameworks like the G7.
• Data Protection Alignment: The emphasis on 'data protection by design and by default' in the context of AI development aligns with global best practices in data privacy, including aspects found in the EU's General Data Protection Regulation (GDPR), even though Taiwan's Personal Data Protection Act has its own specific provisions.

The government's commitment to international alignment is further underscored by the AI Basic Act's call for the establishment of standards and risk categories that are consistent with international norms. This ensures that Taiwan's AI ecosystem, including smart city initiatives like Kaohsiung's, can integrate seamlessly with global technological advancements and regulatory expectations, while also potentially positioning Taiwan as a leader in ethical and responsible AI development on the international stage.

Implementation Timeline

Sources and References