Hong Kong Chief Executive John Lee Ka-chiu has set out a government strategy to expand the use of artificial intelligence across the economy while introducing tighter safeguards against deepfakes, online fraud, risks to minors and other potential harms linked to rapidly developing AI technology.
Speaking after the release of his 2026 Policy Address, Lee said the government would seek to make practical use of AI while keeping its risks under control. The policy programme includes plans to examine new laws covering deepfake related crimes, strengthen mechanisms for removing fraudulent online material and establish a dedicated task force to examine how children use social media and interact with AI.
The approach forms part of a broader seven point AI risk governance strategy contained in the Policy Address. The government said it intends to encourage AI applications in sectors including finance, healthcare, legal services, construction, transport, social welfare, tourism and broadcasting while developing rules covering safety, ethics, liability and the effect of AI on employment.
Lee Says AI Risks Cannot Be Ignored
Lee has acknowledged that AI brings benefits but also creates risks that may be difficult to predict as the technology develops.
The government’s policy document describes AI as a major driver of the digital economy while identifying several areas requiring closer oversight. These include criminal activity involving AI, the responsibilities of users, protection of minors, control of autonomous AI agents and possible changes to the labour market.
The government plans to create a Commissioner for AI within the Digital Policy Office to help coordinate work across government departments. The new role is intended to support the implementation of the government's AI strategy and improve coordination between different policy areas.
Lee has also described his own use of AI in government work. According to the chief executive, AI tools have become part of his daily working routine and assist with tasks such as synthesising publicly available information, developing broader policy concepts and comparing practices in other jurisdictions.
But the use of AI in government comes with clear limits.
Lee has warned against entering sensitive or classified information into publicly available AI systems. He has also stressed that AI generated material should not automatically be treated as accurate and should be checked independently.
That distinction is becoming increasingly important as government departments and businesses look for ways to use generative AI to process information and improve productivity. The Policy Address proposes additional safeguards for AI systems used in public facing government services, including requirements for selected suppliers to comply with government AI guidelines and make safety and ethics declarations about the technology they provide.
Deepfakes and AI Fraud Face Greater Scrutiny
One of the most immediate areas of concern is the misuse of AI for criminal purposes.
The government has asked the Law Reform Commission's Cybercrime Subcommittee to examine existing laws and consider how they should deal with crimes involving deepfake technology. The review will cover the production of pornographic images using deepfakes as well as online scams and fraud in which manipulated or synthetic content is used.
The government has said it will consider new legislation based on the Law Reform Commission's recommendations.
The issue goes beyond manipulated photographs. Advances in generative AI can allow criminals to create convincing audio, video and other material that can be used to impersonate individuals or mislead victims.
Hong Kong is therefore also looking at the removal of fraudulent online content.
The Financial Services and the Treasury Bureau plans to conduct a public consultation later this year on changes to the legal framework that would encourage technology and telecommunications companies to strengthen their ability to detect and remove fraudulent material, including scams generated or assisted by AI.
The proposed approach means that the government is looking at both criminal law and mechanisms for taking harmful material off the internet.
That work could become particularly significant for financial institutions and online platforms, where fraudsters can use increasingly sophisticated synthetic content to target victims.
Government Wants Wider AI Adoption
At the same time, Hong Kong is not planning to restrict AI development simply because of the risks.
The Policy Address sets out an extensive programme for applying AI across different parts of the economy and public sector. The government says it wants to promote AI adoption while maintaining security and encouraging responsible use.
In financial services, the Hong Kong Monetary Authority has expanded its generative AI sandbox into a cross sector programme covering areas beyond banking, including securities, wealth management, insurance and mandatory provident fund services. The initiative will allow participating organisations to test selected AI applications in a controlled setting.
Healthcare is another area identified for wider adoption. The Health Bureau will coordinate with relevant organisations, including the Department of Health and Hospital Authority, on the orderly introduction and use of AI technologies.
The legal sector will receive a short term funding scheme to encourage wider use of LawTech, including AI applications, while the Housing Authority plans to expand an AI based platform that can analyse factors such as terrain, views, wind and sunlight when assessing building layouts.
The government has also identified transport, social welfare, tourism and broadcasting as areas where AI applications can be expanded.
New Measures for Children and Social Media
Protecting minors is another major component of the government's AI risk programme.
Under the 2026 Values Education Curriculum Framework, the Education Bureau will support schools in teaching students how to use AI tools and resources appropriately. The focus will include responsible use, values and attitudes surrounding new technology.
A dedicated task force on minors' use of social media will also be established under a working group led by the Department of Justice.
The task force will examine risks associated with children's use of social media, including interactions involving AI. It will consider whether more targeted measures are needed, including possible legislative or administrative changes.
Officials have said the government will examine approaches adopted in other jurisdictions before deciding what measures may be appropriate for Hong Kong.
The review comes as children increasingly encounter AI powered tools through social media, education platforms and online services. The government is therefore considering the issue not simply as a question of internet use, but as part of a wider discussion about how young users interact with AI systems.
Government Examines AI Ethics and Liability
Hong Kong's proposed AI governance framework also extends beyond fraud and child protection.
The government plans to introduce pilot requirements for suppliers bidding on selected public facing AI service projects. Those companies will need to comply with AI guidelines issued by the Digital Policy Office and provide declarations concerning the safety and ethical compliance of the technology they propose to use.
If the pilot scheme produces satisfactory results, the requirements could gradually be extended across government departments.
Officials will also examine who should bear legal responsibility when an AI product causes an accident or damage.
A working group led by the Department of Justice will review whether existing laws are sufficient to deal with such cases and consider whether new legislative or regulatory measures are required. Possible approaches include codes of practice, codes of conduct and formal guidelines.
Another part of the strategy concerns AI agents, which can independently combine and execute multiple tasks.
The Digital Policy Office is already piloting AI agents within the government and has incorporated relevant provisions into its AI adoption guidance for departments. Based on the experience gained through those trials, the government plans to develop safety management guidelines for AI agents for wider use next year.
AI Could Reshape Hong Kong's Workforce
The government is also examining what wider AI adoption could mean for employment.
The Labour and Welfare Bureau is due to publish an updated manpower projection examining the effects of AI adoption on Hong Kong's labour market. The government says it will use the findings to develop measures including training, upskilling and reskilling support.
Separately, Hong Kong plans to launch an 18 month AI upskilling campaign in the first half of 2027. The programme will include a free online course focused on workplace applications of AI, with participants who complete it able to enrol in free short term advanced courses covering different industries. The government expects about 40,000 employees to benefit.
This creates two parallel strands in Hong Kong's AI policy. The first is to accelerate adoption by companies, public bodies and professionals. The second is to prepare workers and institutions for the disruption that broader use of AI could bring.
A Policy Built Around Adoption and Control
The government's latest plans show that Hong Kong intends to expand AI use rather than wait for the technology to mature before acting.
The Policy Address combines industry support with measures aimed at limiting criminal misuse, protecting children, establishing ethical standards, clarifying legal responsibility and preparing workers for changes in the labour market.
Lee's own use of AI illustrates the government's approach. Officials are being encouraged to use the technology for research, analysis and productivity, but sensitive information must remain protected and AI generated material must still be independently checked.
For businesses, schools, government departments and ordinary users, the next stage will depend on how these proposed measures are translated into legislation, guidelines and day to day practice.
As Hong Kong moves to expand AI across finance, healthcare, transport, construction and public services, one question will become increasingly important: how can the city capture the productivity gains of AI while ensuring that stronger safeguards keep pace with the technology's risks?