The Zhitong Finance App learned that the Secretary for Innovation, Technology and Industry of Hong Kong, Sun Dong, said in response to lawmakers' questions that the government has been actively deploying relevant digital infrastructure to seize the opportunities brought by the rapid development of AI. Currently, the overall computing power of Hong Kong has reached 5,000 trillion floating point operations per second (5,000 PFLOPS), of which the Cyberport Artificial Intelligence Supercomputing Center accounts for 60%. Furthermore, the Shaling Data Park, which has already entered the construction stage, is expected to provide 180,000 PFLOPS of computing power by 2032, which is equivalent to 36 times the current computing power of Hong Kong, providing strong support for AI computing needs, including word processing.
Sun Dong pointed out that a word element is the basic unit for AI systems when processing text, and is commonly used to reflect data processing volume and related computing power requirements. The Government will keep a close eye on the latest developments in global and local AI, and continue to optimize various digital infrastructure.
He also pointed out that while promoting innovative applications of AI, it is necessary to effectively manage technical risks and ensure that operations meet safety and ethical standards. The Digital Office has issued a number of key guidelines, including the “Ethical Framework for Artificial Intelligence” and the “Hong Kong Generative Artificial Intelligence Technology and Application Guidelines” for generative AI technology, which apply within the government and are also used as reference by public institutions and all sectors of society.
The Government has formulated relevant regulations for departmental applications of AI, protection of personal data privacy and information security. In addition to the overall guidelines mentioned above, the Digital Office also issued the “Government Departmental Artificial Intelligence Application Guidelines” in February 2026 to ensure that content (including videos and images) generated by AI tools must be verified for authenticity by official or trusted sources, and prohibited from entering confidential or personal data into external or public AI platforms, and not directing the generation of content that infringes copyright or intellectual property rights.
The Government IT Security Policy and Guidelines also set standards for account registration, management and system operation, including encrypting transmitted and stored data, and require departments to conduct regular security risk assessments and audits of their information systems and related data.
In order to encourage departments to explore AI applications under safe and controllable conditions, promote process optimization, and reduce the time and administrative costs of independent procurement systems for various departments, the Digital Office has launched the “Artificial Intelligence Solution Catalogue” and AI toolbox, bringing together more than 260 AI solutions (such as text summarization, automatic translation, data analysis, and document assisted writing) that have been technically reviewed and applied to government scenarios, and list their applicable scenarios, technical architecture and restrictions.
Furthermore, the Government continues to strengthen talent training to enhance civil servants' digital thinking and AI application capabilities. The Digital Office and the Civil Service College launched a hierarchical online AI training course, covering topics such as basic knowledge, ethical security and tool applications, to cutting-edge technology trends and regional development, and conducted tests to evaluate and consolidate learning results; it also regularly organizes a series of special lectures on “Innovation and Technology” leadership to strengthen management's digital governance capabilities. In 2025/2026, the Digital Office arranged about 130 training courses related to digital technology to promote the integration of theory and practice, and cultivate a civil service that makes good use of technology.
Whether in guidance or training, the government emphasizes the importance of “people in the ring”. Before publishing any AI-assisted content, government departments must also go through manual review and verification by personnel to ensure that the content is true and accurate, and must not rely entirely on AI output.