Sun Dong: The development of AI in Hong Kong emphasizes both innovation and security, actively building a healthy AI ecosystem.

In response to the questions raised by legislators regarding artificial intelligence, Professor Wang Dong, the Secretary for Innovation and Technology of Hong Kong, replied that Hong Kong has been following the strategy outlined in the "Hong Kong Blueprint for Innovation and Technology Development" to actively build a local AI ecosystem and promote the vibrant and safe development of AI in Hong Kong. The work in this area is ongoing and deepening. The recently established Hong Kong AI Research Institute will begin operations in the second half of this year, leading and coordinating research and development in AI, the transfer of results, and the exploration of application scenarios. In terms of infrastructure, Hong Kong's overall computing power has reached 5000 petaflops per second (PFLOPS), and with the construction of the Sha Tin Data Center underway, it is expected to provide 180,000 PFLOPS of computing power by 2032, equivalent to 36 times the current computing power of Hong Kong. On the legal framework front, in order to review whether the laws in different policy areas can support technological developments including AI, the Secretary for Justice convened a supervisory committee meeting on March 6 this year for the establishment of a cross-departmental working group to review the legal framework needed to support a broader application of AI. The establishment of a working group consisting of representatives from different decision-making bodies and government departments is essential for addressing the issues arising from the rapid development and widespread application of AI. He mentioned that the government attaches great importance to the governance and risk prevention of AI applications. In terms of governance, the Office of the Government Chief Information Officer (OGCIO) has developed the "AI Ethics Framework" and the "Hong Kong Generative AI Technology and Application Guidelines" to provide guidance for the development and application of AI technologies, to identify and manage potential risks and considerations for AI projects (such as the security and robustness of systems, personal privacy, data security, and management), and to provide practical guidance for technology developers, service providers, and users. For example, the "AI Ethics Framework" suggests that project managers conduct impact assessments of AI applications, consider ethical standards, roles, and responsibilities related to AI according to risk classification criteria, and determine which projects require review. Especially high-risk applications should have a "human-in-the-loop" supervision mechanism to reduce the risks associated with using AI. On information security, the OGCIO has formulated a comprehensive set of "Government Information Technology Security Policy and Guidelines," which include scenarios for the application of AI tools. The relevant guidelines have been uploaded to the government website, and individual organizations can adopt suitable information technology security risk management measures to enhance their awareness and capabilities in information security and risk prevention related to AI. The government will consider expanding the scope of application of these guidelines to organizations and individuals outside the government based on actual circumstances. Professor Wang mentioned that in recent years, generative or agent-based AI intelligences have become more prevalent, but certain applications may involve issues such as excessive permissions, data leakage, and system intrusions. The OGCIO has reminded various government departments and the public of security risks and provided relevant security guidelines, such as not installing AI intelligence agents on computers connected to the government's internal network, conducting risk assessments before installing different software, including the need for using the software/product, product characteristics and past records, security risks to the department, etc. It also suggests that users must take sufficient security measures when deploying and using AI intelligence agents, such as strict isolation of the operating environment, strengthening credential management, and staying attentive to official security updates to mitigate related risks. To enhance the public's awareness of the security aspects of AI applications, the OGCIO supports the Hong Kong Internet Registration Corporation Limited in continuously updating AI-related workplace security application courses through the network security staff training platform, covering workplace application risk alerts and codes of conduct. In addition, the 2026-27 Budget announced an allocation of HK$50 million to enhance the public's understanding and capabilities in using AI responsibly and foster a culture of AI literacy. The Hong Kong government will continue to monitor the latest developments in AI and emphasize innovation and security while promoting the widespread application of AI. It will update guidelines and governance frameworks as needed to establish a healthy and orderly AI ecosystem.