Visitors check out Huawei's Pangu-Weather, a powerful AI-based weather forecasting system, during the World Artificial Intelligence Conference 2023 in Shanghai on July 6. [Photo/Xinhua]

Country to boost policies, develop national standards for LLM technology

China is actively encouraging artificial intelligence companies as well as research institutes to build large language models, as part of a broader effort to boost the country's prowess in the critical technology behind ChatGPT, a text-generating AI chatbot developed by US-based tech firm Open-AI that went viral across the world after it was launched late last year.

Industry experts believe the move will be a game changer in the deepening tech rivalry between the United States and China, the world's two largest economies.

On Tuesday, the Beijing Municipal Bureau of Economy and Information Technology told China Daily that Beijing will offer over 40 million yuan ($5.6 million) in subsidies for the first batch to support companies involved in LLM training.

"The US continues to move forward. If China cannot make its own universal large language model in the second half of this year, we might not even be able to see the 'taillights of cars' in the future," said Yu Linwei, deputy head of Shanghai's Xuhui district, at the World Artificial Intelligence Conference in Shanghai last week.

The China Electronics Standardization Institute, which is under the Ministry of Industry and Information Technology, the country's top regulator for industries, said during the conference that a national standard for LLMs would be implemented soon. It has enlisted internet search leader Baidu Inc, telecom giant Huawei Technologies Co, cybersecurity firm 360 Security Technology and e-commerce company Alibaba Group, to lead in the development of the standard.

The authorities have been planning to standardize LLMs in the country since the ChatGPT frenzy last year, said an industry veteran who is close to the institute, asking not to be identified by name. As this is guided by an AI working group of the National Information Security Standardization Technical Committee set up in 2020, almost all leading internet firms are involved, the source said.

"It means that the names of the companies announced this time in Shanghai are only the first batch. A large number of tech companies and research institutes have already joined in the making of such standards," the source added.

A company executive who is familiar with the matter said that Xiaomi Corp; CloudWalk, an AI company backed by the Chinese Academy of Sciences; SenseTime and iFlytek are all member companies of this working group.

Beijing has the largest number of such models, 38, followed by Shanghai as well as Guangdong and Zhejiang provinces, the report said.

China's capital city has launched a package of supportive policies to drive the innovative development of generative AI, said Jiang Guangzhi, deputy director of the Beijing Municipal Bureau of Economy and Information Technology, at the Global Digital Economic Conference in Beijing last week.

In addition to standard-setting, the city has encouraged local companies and government departments to "purchase and use safe LLMs and related services", Jiang said.

He said Beijing also plans to launch policies related to computing power that will help small- and medium-sized enterprises conduct LLM training and applications amid the ChatGPT boom.