Wuhan Unveils High-Quality Dataset Plan, Targets 200 Datasets and RMB 200 Billion AI Industry by 2028

Wuhan recently issued the Action Plan for Building High-Quality Industry Datasets in Wuhan (2026-2028). The plan proposes that by 2028 the city will have cumulatively built more than 200 high-quality industry datasets, formed a batch of standardized corpus resources that meet large-model training needs, and helped the artificial intelligence industry scale exceed RMB 200 billion (2000 hundred million yuan).

The publishing industry has rich corpus resources. The image-text resource libraries formed by publishing companies through book digitization are important datasets for AI model training. Their copyright and IP resource advantages give them a key foundation in AI model research and development.

CITIC Press holds a large number of high-quality copyright resources. Its self-developed Kuafu AI digital publishing platform runs through the entire book publishing process, reducing costs and improving efficiency. The company's Knowledge Essence AI knowledge base has been connected to corporate learning platforms.

Guomai Culture has accumulated massive text corpora over years in the general-interest book publishing industry. After screening and refinement by R&D personnel, the data quality is higher, allowing models to learn standardized language habits and proofreading methods and rules.