In China, local authorities have approved the launch of the first batch of large language models (LLM) this year.
LLM is a technology that is used to train ChatGPT and similar services. Currently, there is an increasing number of local technology companies in China that offer various solutions for the implementation of artificial intelligence in sundry spheres of activity.
For commercial use, as reported by the media of the Asian country and individual companies, 14 language models were approved, including the development of smartphone manufacturing giant Xiaomi, machine intelligence specialist Beijing Fourth Paradigm Technology Co, and Beijing firm tech unicorn 01.AI, founded by venture capitalist Lee Kai-fu.
LLMs, which were approved for launch in January, include industry-specific solutions. This feature is the lineament of the current batch of models. Previous approvals by the Chinese authorities concerned general configurations of artificial intelligence. In this case, functional versatility is implied. Currently, in China, the development and regulation of machine intelligence are aimed at promoting advanced technology to improve the efficiency of enterprises.
Last August, the public launch of generative artificial intelligence services was approved for the first time in an Asian country. A month earlier, the China Electronic Standardisation Institute was established to implement a national standard for language models.
In January, approval to launch LLM Pinshang and Moxiaoxian was received by the startup Frontis.AI. This company was founded in 2021 by a former senior vice president of JD.com Zhou Bowen. The mentioned models are offered to firms, that specialize in the implementation of consumer goods manufacturing activities, industry insights, and support for innovation in the sphere of products and marketing. The startup intends to start providing more customized services to corporate clients over time. A statement of this plan was published last week on the company’s official account registered on the WeChat virtual platform.
The industry licenses of ThreatBook, a firm specializing in cybersecurity solutions, and XinYi Tech, a video solution provider, were also approved.
The XGPT model, developed by ThreatBook specialists, is the first LLM designed to provide a high level of security in China’s cyberspace. The company said that its development will help to receive timely and accurate analysis of potential threats.
LLM, developed by XinYi Tech specialists, is the first Chinese AI model that specializes in creating tools for generating and processing video materials. The company stated that its product is intended for such spheres of activity as advertising, media, education, and e-commerce.
There is no information yet about other LLMs that have been approved by the Chinese authorities.
Apps with artificial intelligence, which were approved in January, include an AI-based resume-processing tool created by a digital recruitment platform called Zhaopin. The launch of chatbots by Beijing Zhidemai Tech, an e-commerce service provider, and iReader Technology, a provider of mobile books, magazines, and comics, was also allowed.
The number of LLMs and related applications approved in China currently stands at over 40.
Robin Li Yanhong, founder, chairman, and CEO of the Internet search engine and artificial intelligence giant Baidu, says that the hype about language models was a huge waste of resources. He called on the Chinese government to take more ambitious initiatives to bring more services to the market that are similar to ChatGPT in terms of functionality. According to him, 238 LLMs have already been developed in China as of October, most of which have not yet been approved.
As we have reported earlier, Baidu’s AI Chatbot to Power Samsung’s New Galaxy S24 Smartphones.
