Can Hong Kong’s ChatGPT-style Tool Overcome OpenAI’s Access Blocks?

Hong Kong Tests New ChatGPT-style AI Tool for Government Use Amid OpenAI Restrictions

The Hong Kong government is piloting its own ChatGPT-style AI tool for civil servants, with future plans to make it accessible to the public, according to the city’s innovation minister. This comes after OpenAI tightened access restrictions for Hong Kong and other unsupported regions.

Secretary for Innovation, Technology and Industry, Sun Dong, announced on a radio show that his bureau is testing an AI program known in Chinese as the “document assistance application for civil servants” to enhance its capabilities. The goal is to roll out the program across the entire government within this year.

Developed by a generative AI research and development center led by the Hong Kong University of Science and Technology, in collaboration with other universities, the AI tool is expected to eventually offer features like graphics and video design. However, the extent to which it will match ChatGPT’s capabilities remains uncertain.

The Associated Press received no response from Sun’s bureau regarding the model’s functions. On the radio show, Sun emphasized the collaboration between industry players and the government in the tool’s ongoing development.

“Given Hong Kong’s current situation, attracting major companies like Microsoft and Google to fund such projects is challenging, so the government has taken the initiative,” Sun explained.

As Beijing and Washington vie for AI dominance, with China aiming to lead globally by 2030, Hong Kong and Macao remain excluded from OpenAI’s list of “supported countries and territories.” OpenAI has not disclosed reasons for these exclusions but has stated that accounts from these regions may be blocked from accessing its services. An email to some users on OpenAI’s forum indicated that stricter measures would start on July 9, though no further explanations were given.

Francis Fong, honorary president of the Hong Kong Information Technology Federation, expressed uncertainty about the Hong Kong AI program’s potential to rival ChatGPT but was optimistic about its development with local AI companies’ input. He noted that a locally developed AI might better address regional languages and issues but also hinted at the likelihood of it being “politically correct.”

Despite China’s internet censorship via its firewall, tools like virtual private networks allow determined users to access foreign websites and applications, including ChatGPT. Chinese tech giants such as Alibaba and Baidu have introduced their own Chinese-language AI models similar to ChatGPT, albeit within the constraints of China’s censorship regulations.

In May, China’s cyberspace academy revealed an AI chatbot trained on President Xi Jinping’s doctrine, highlighting the ideological limits within which Chinese AI models must operate. Concurrently, SenseTime launched SenseChat for Hong Kong users, primarily Cantonese speakers. However, politically sensitive queries, such as those about the Tiananmen Square crackdown in 1989 or the 2019 Hong Kong protests, yielded no results.

The 1989 crackdown saw Chinese troops firing on pro-democracy protesters, resulting in numerous deaths—a topic still taboo in mainland China. The 2019 Hong Kong protests, sparked by controversial legislation, evolved into an anti-government movement posing a significant challenge to Beijing since the city’s 1997 return to Chinese sovereignty.