China Aims for 50 AI Standards by 2026, Covering Chips, Applications, and Safety
China intends to produce at least 50 AI standards by 2026, including topics such as LLM training, safety, governance, and industrial uses. These standards are projected to cover more than 1,000 Chinese technology enterprises. China intends to contribute in the development of twenty international AI standards.
The Ministry of Industry and Information Technology (MIIT) has released a draft policy that outlines proposed standards for various aspects of AI, such as large language model (LLM) training, safety measures, governance, industrial applications, software, computing systems, data centres, and semiconductor requirements.
According to the document, these criteria will apply to more than 1,000 Chinese technology enterprises. Furthermore, China intends to contribute in the development of at least 20 international AI standards. This approach is consistent with China's resolution at the United states, which advocates for a "free, open, inclusive, and non-discriminatory" commercial environment for AI development across states.
The MIIT thinks that AI is a fundamental and strategic technology that will propel the next wave of technological revolution and industrial change. China intends to fundamentally alter industrial production and economic development patterns by incorporating AI into its economy. The organisation underlines that artificial intelligence will play a critical role in improving China's manufacturing and internet capacities.
In contrast to traditional command-and-control regulation, the MIIT's draft policy takes a pro-market and soft-law approach to guiding and promoting the growth of China's AI industry. This method is viewed as innovative and market-friendly, enabling and encouraging the evolution of technology and its ecosystem, which benefits not only the AI industry but also other sectors.
The draft policy specifies 12 important technologies for the AI supply chain, including LLMs, natural-language processing, computer vision, and machine learning. These technologies provide as the cornerstone for China's AI industry chain, which is divided into four layers: computer power, algorithms, and data for LLM training; the framework; the model; and applications.
Earlier this year, Alibaba Group Holdings chairman Joe Tsai admitted that China trails the United States by two years in the global competition for AI leadership. Washington's technology export curbs have posed hurdles for mainland businesses.
China plans to establish at least 50 AI standards by 2026, covering areas such as LLM training, safety, governance, and industrial applications.
Over 1,000 Chinese technology companies are expected to be covered by these standards.
China aims to participate in the establishment of 20 international AI standards.
Source: SCMP