China shifts from general AI models to industry-focused systems
At a conference in Hong Kong, experts revealed that China is pivoting from general-purpose language models toward industrial AI systems specialized in areas such as factory data and financial records.
China's artificial intelligence strategy is evolving beyond general large language models (LLMs) toward specialized systems focused on solving real-world industrial problems.
Speaking at the Hong Kong International Computer Conference, Associate Professor Cui Peng from Tsinghua University stated that large language models struggle to capture patterns in structured numerical data. Cui noted that expert systems capable of analyzing data such as factory sensors and financial records will constitute the next phase of China's AI development.
Cui's research team, in collaboration with Wenzhun Intelligence, has developed a large data model called 'LimiX', trained with synthetic data to increase efficiency in industrial and scientific environments. Nie Jun Yu from Huawei Technologies emphasized the transition to industry-wide systems coordinated by local governments in provinces such as Zhejiang and Guangdong, which position AI as a 'digital public good'.
Hong Kong's Secretary for Innovation, Technology and Industry, Sun Dong, announced that the city aims to accelerate industrial adoption in line with Beijing's 'AI plus' initiative by expanding its AI computing infrastructure.
Sources
- scmp.com · Link