Hong Kong should ban AI tools posing threats to human safety: policy office
Hong Kong's digital policy body has called for a ban on generative artificial intelligence (AI) systems that pose threats to human safety and extensive supervision of such software if it must be used in critical infrastructure, as part of new guidelines.
Advertisement
Speaking at the second day of the World Internet Conference Asia-Pacific Summit on Tuesday, Commissioner for Digital Policy Tony Wong Chi-kwong announced the city had published the new technical and application guidelines for generative AI tools.
'Its primary objective is to balance AI innovation application and responsibility, thereby constructing a governance framework tailored to the Hong Kong context, with local characteristics for all stakeholders in the AI ecosystem,' he said in a speech.
'We hope the guidelines can facilitate the industry and the public in developing and applying generative AI in a safe and responsible manner, while encouraging innovative application of AI, mitigating risk and fostering the widespread adoption of generative AI in Hong Kong.'
The office earlier commissioned the Hong Kong Generative AI Research and Development Centre, which falls under the government's InnoHK innovation programme, to research the topic and propose guidelines.
Advertisement
The centre reviewed the technology's application and the positions taken by various jurisdictions, in addition to gathering feedback from the innovation and technology industry to compile the guidelines covering technologies, developers, service providers and users.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
RTHK
24-05-2025
- RTHK
Digital Policy Office to review AI guidelines annually
Digital Policy Office to review AI guidelines annually Tony Wong said each industry should have risk management measures that cover artificial intelligence use. Photo: RTHK Digital Policy Commissioner Tony Wong on Saturday said his office plans to review newly-published guidelines on artificial intelligence use at least once a year. Last month, the Digital Policy Office released a document on generative artificial intelligence aimed at developers, service providers, and users. Speaking on an RTHK radio programme, Wong said AI technology constantly changes, so guidelines must be regularly updated. "We don't have a deadline. But I think we should update each year, or even shorter," he said. "That's because technological development is so rapid. If there's a new application or risk, the guiding principles may need to be changed." He said each industry should have risk management measures that cover artificial intelligence use. Government departments have adopted the AI document application HKPilot, launched last year by the Hong Kong Generative AI Research and Development Centre. Wong said that more than 15,000 staff have tried the application. Its functions include summarising and analysing documents. Wong also said the government's "iAm Smart" app, which is used as a platform for government departments, has seen over 3.5 million registrations. He said he hoped this figure would rise to four million by the end of the year when more functions would be available.
South China Morning Post
03-05-2025
- South China Morning Post
AI guidelines support innovation and target those who exploit tech in Hong Kong
Artificial intelligence (AI) has been advancing at an astonishing pace, raising concerns about how the technology will change the world – and how soon. It is good to see Hong Kong taking timely action to craft policies to help ensure AI tools are beneficial rather than threats to human safety. Advertisement Commissioner for Digital Policy Tony Wong Chi-kwong has called for a ban on generative AI systems that may pose such threats, and for extensive supervision of AI software used in critical infrastructure. Wong made his call last month at the World Internet Conference Asia-Pacific Summit. Hosting 1,000 local and overseas participants for the first such event held outside the mainland, the city showed how it can serve as a bridge and two-way platform linking China with the rest of the world in AI development. Wong showcased the city's new technical and application guidelines for generative AI tools to 'balance AI innovation, application and responsibility' and build a governance framework tailored for local characteristics and all stakeholders. The guidelines aim to facilitate the 'safe and responsible' development and application of generative AI. Advertisement The guidelines had input from the Hong Kong Generative AI Research and Development Centre, which studied how other jurisdictions have applied the technology.
South China Morning Post
25-04-2025
- South China Morning Post
Hong Kong to launch local ChatGPT-style AI tool in second half of year
Hong Kong is set to launch a locally developed ChatGPT-style artificial intelligence (AI) tool powered by DeepSeek in the second half of the year, with the platform featuring elements of the city and being able to answer 'all kinds of questions', the innovation chief has said. Advertisement Secretary for Innovation, Technology and Industry Sun Dong gave an update on the launch timeline on Friday during a Beijing-organised media tour to promote the Greater Bay Area. The new tool, named 'HKGAI V1', was developed by the Hong Kong Generative AI Research and Development Centre (HKGAI) under the government's InnoHK innovation programme. The tool is capable of delivering instant responses to user queries through a chatbot interface. Users can input commands to access information or generate specific content such as travel itineraries, meeting notes, music and videos. 'The system can answer all kinds of questions,' Sun said, adding that it would feature local characteristics reflecting the values of Hong Kong, including the 'one country, two systems' governing principle. Advertisement He also expressed his hope that the tool could serve the Chinese community overseas. HKGAI V1 is the first in the world to be based on the DeepSeek full-parameter fine-tuning research model.
