Tech war: OpenAI to further block access by mainland China, Hong Kong-based developers

ChatGPT creator OpenAI is tightening measures to block attempts from “unsupported countries and territories”, including mainland China and Hong Kong, to access its generative artificial intelligence (AI) services via application programming interfaces (APIs), according to a notice sent to Chinese developers, deepening the AI divide between China and the US.

While Microsoft-backed OpenAI has rolled out its AI services in over 160 countries, the products remain officially unavailable in mainland China and Hong Kong. Users there have been turning to virtual private networks or third-party apps to access ChatGPT, while developers need to use proxies and outbound servers to bypass restrictions.

OpenAI’s move, which is set to go into effect on July 9, could affect Chinese companies developing their own services based on OpenAI’s large language models (LLMs) – the technology that underpins ChatGPT and similar generative AI applications – according to some experts.

The latest move by OpenAI essentially represents the US further restricting China’s access to advanced AI technology, according to an industry insider who declined to be named for discussing politically sensitive matters.

A number of AI start-ups in China are building apps based on OpenAI’s large models, which also generate revenue for OpenAI, the person said, adding that if OpenAI strengthens its regulations, Chinese developers will have to turn to local alternatives.

“We are rolling out a special migration plan for shifting to home-grown LLMs,” the Beijing-based start-up said on Tuesday, dangling perks such as 150 million free tokens, as well as tailored training courses to support such moves.

Rapid development in AI has opened a new front in the US-China tech war. The Biden administration is limiting China’s access to advanced semiconductors and restricting American investments in the mainland’s AI sector. Washington is also reportedly planning new regulations to curb the export of advanced AI models.

OpenAI has been mindful of foreign misuse of its services.

In May, the San Francisco-based company said it had disrupted five “covert influence operations”, including networks in China, Russia, Iran and Israel that used its AI products to manipulate public opinion or shape political outcomes while obscuring their true identities.

In a separate move, OpenAI said in December it suspended TikTok owner ByteDance’s access to its services, citing policy violations. The Chinese company later said that a small group of its engineers had used OpenAI’s API service in an experimental model not intended for public launch.

OpenAI – which provides some of the technology behind Apple Intelligence, a suite of AI features recently unveiled by the iPhone maker – is not expected to support Apple’s new AI functions in China.

Apple is reportedly seeking local partners to provide services that adhere to Beijing’s various AI regulations, including those laid out by the China Electronic Standardisation Institute last year to enforce national LLM standards.

Chinese tech companies, which have introduced more than 200 self-developed LLMs, are still racing to catch up with their global peers in AI.

Some of the front runners include Big Tech companies such as Baidu, ByteDance and South China Morning Post owner Alibaba Group Holding, as well as start-ups including Zhipu AI, Baichuan, Minimax and Moonshot AI.

Additional reporting by Ben Jiang