US advisory body: China's data dominance may offer AI advantage

The commission highlighted that while U.S. AI companies have largely utilized publicly available internet data for training language models, China is systematically gathering proprietary enterprise, operational, and physical-world data. This type of data is essential for developing sophisticated AI tools for business applications, autonomous vehicles, and robotics.

Specifically, China's robust advanced manufacturing and industrial robotics sectors are noted as providing a rich source of high-quality data for embodied AI systems, which operate in the physical world. This could give China a competitive edge in developing commercial and military robotics software.

Mike Kuiken, vice chair of the commission, stated that China has effectively consolidated, labeled, and refined data, making it rapidly available to relevant entities. He also noted that this strategy benefits China's innovation ecosystem and enhances the control of the Chinese Communist Party.

In recent years, China has implemented stricter regulations on cross-border data transfers, requiring foreign firms to localize their Chinese operations and separate their data. While some exemptions exist, U.S. companies operating in China face risks related to these stringent data and cyber governance regimes.

China has elevated data to a core factor of production, establishing the National Data Administration in 2023 to manage data standardization and classification, aiming to create a unified national market for data trading. The country is pursuing the commercialization and deployment of data across industries like manufacturing, transport, finance, and healthcare, with Chinese firms beginning to list datasets on regional exchanges.

As a key recommendation, the commission urged the U.S. Congress to consider developing a national data strategy that treats data as a valuable economic asset.