China’s ‘two sessions’ 2023: ChatGPT-like artificial intelligence is ‘difficult to achieve’, China’s tech minister says

Science and Technology Minister Wang Zhigang raised ethical concerns and said the country must ‘wait and see’ when it comes to catching up with ChatGPT
The hit chatbot from OpenAI has become a hot topic in China, but domestic tech giants must overcome censorship and chip restrictions to create rival products

Two Sessions 2023 (Lianghui)

Xinmei Shen

Published: 10:00pm, 5 Mar 2023

Why you can trust SCMP

China has some work to do in catching up to the hit artificial intelligence product ChatGPT, developed by San Francisco-based start-up OpenAI, the head of the country’s Ministry of Science and Technology said, as technological self-reliance remains a top priority for the central government amid heated trade tensions with the US.

ChatGPT has advantages in delivering results in real time, which is “very difficult to achieve”, Science and Technology Minister Wang Zhigang said at a press conference on the first day of the annual National People’s Congress (NPC) on Sunday.

China over the years has done a lot of planning and research in the fields of natural language processing and natural-language understanding, and has made some achievements, according to Wang. But for China to achieve the kind of results seen by OpenAI, the country needs to “wait and see”, he said.

China’s ChatGPT-like AI should have age limits, ‘two sessions’ delegate says

Wang also raised the issue of AI ethics on Sunday, saying companies need to be mindful of such concerns.

The large language models that produce sophisticated bots like ChatGPT can result in unpredictable, and often inaccurate, responses. Questions have been raised about whether China’s tight internet censorship could allow for the development of a product like ChatGPT.

The explosive popularity of ChatGPT since it was launched in November has pushed generative AI into the global spotlight, spurring many conversations in China about the direction of the domestic industry.

The chatbot is a topic of discussion at this year’s “two sessions”, the country’s largest annual political gathering, where the NPC and the National Committee of the Chinese People’s Political Consultative Conference (CPPCC) unveil national policy plans for the coming year.

Minister of Science and Technology Wang Zhigang gives an interview after the opening meeting of the first session of the 14th National People’s Congress (NPC) in Beijing on March 5, 2023. Photo: Xinhua

Ahead of the event, Zhou Hongyi, the founder and chairman of 360 Security Technology, issued a proposal for China to nurture its own ChatGPT-like technology. Zhou Yuan, founder and CEO of question-and-answer platform Zhihu, also said in a state media interview that minors’ access to such AI should be limited.

Both executives are attending the “two sessions” as CPPCC delegates.

Many of China’s largest tech firms have already promised to develop their own alternatives to ChatGPT, which despite not being officially available in China has gone viral in the country through third-party apps and services.

Most notably, Chinese internet search giant Baidu announced last month that it will unveil its ChatGPT rival Ernie Bot as soon as this month. On-demand service giant Meituan’s co-founder Wang Huiwen pledged to invest US$50 million in a ChatGPT-like project with a valuation of US$200 million. E-commerce giant Alibaba Group Holding, owner of the South China Morning Post, said its research institute Damo Academy is developing a similar tool.

In addition to overcoming domestic censorship challenges, companies will also need an immense amount of computing power at a time when it is becoming harder to procure in China. US restrictions on exporting to China powerful graphics processing units (GPUs) like Nvidia’s A100, which are integral to machine learning, could slow national AI development goals.

Wang has previously commented on the ethics of AI development as it relates to ChatGPT. “For any new technology, including AI technology, our country has adopted some corresponding measures in regard to ethics,” Wang said during a press conference last month.

In 2021, the Ministry of Science and Technology drew up its first set of ethical guidelines governing AI in a document titled “New Generation Artificial Intelligence Ethics Specifications”.

Among a range of requirements, the guidelines stated that humans should have full decision-making power over AI and have the right to choose whether to accept related services, exit an interaction with an AI system or discontinue its operation at any time.