New batch of large language models registered, underscoring expedition of vertical AI application
Published: Jan 30, 2024 08:20 PM



A new batch of artificial intelligence (AI) large language models (LLMs) was recently incorporated into regulations, underscoring the country's dedication to both expediting the implementation of vertical AI models and effectively governing AI development.

The latest batch that passed the interim regulation on the management of generative AI services includes 14 domestic LLMs from 13 Chinese enterprises, Beijing Business Today reported on Monday. 

In July, China released an interim regulation on the management of generative AI services, which went into effect on August 15, 2023. It put forth measures on boosting generative AI technology, while stipulating basic norms for generative AI service providers.

On the list are an AI question answering bot produced by Beijing Zhidemai Technology and the AI model business shooting tool "moxiaoxian" by Beijing-based AI startup Frontis, as well as a large language model by Lingyi Wanwu Information Technology Company, the AI dialogue WeChat mini program "yueailiao" by Zhangyue Technology and the first large AI model focusing on the video field by Xinyi Technology.

The registration for LLMs is of great significance for protecting users' rights and interests and preventing technological abuse as the country is witnessing an AI boom, industry observers noted.

This batch of registered domestic models cover a number of enterprises and a variety of technologies, showing the diversity and innovative capabilities of Chinese AI enterprises, Guo Tao, an angel investor and veteran AI expert, told the Global Times on Tuesday. Guo said these models have made breakthroughs in algorithm efficiency, data processing capacity and application scenarios, and can better serve the needs of specific industries. 

The AI expert said that the models' registration for the interim regulation on the management of generative AI services signals a fresh batch of domestic models will accelerate their commercialization.  

More than 40 AI LLM products have been approved for generative AI services in China. At the same time, more than 240 AI technology projects have registered as domestic deep synthesis service algorithms, Chinese digital information outlet TMT post reported. 

Industry analysts also say that the continual approval of new LLMs into regulations, on the one hand, reflects the attention and support of regulatory authorities have when it comes to the development of these models, and on the other hand, it also shows the determination of regulatory authorities to manage and standardize LLMs. 

Guo said that he is aware that some small products have been released without registration due to their decreased security risks, but he calls for all AI products, regardless of size, to undergo strict registration and review to ensure their safety and compliance.

The year 2024 is expected to bring promising developments in China's AI field, with a particular focus on the applications of generative AI technology, improvements in regulatory policies to ensure the responsible use of AI technology and data privacy protection, Wang Peng, an associate research fellow from the Beijing Academy of Social Science, told the Global Times in an earlier interview.