China’s New Trade Secret Regulations: Protection Extended to Data and Algorithms
China has brought into force a new set of trade secret regulations that, for the first time, explicitly include data and algorithms within the scope of protectable technical information. Issued by the State Administration for Market Regulation (SAMR), the Provisions on the Protection of Trade Secrets represent the most significant overhaul of China’s trade secret regime in nearly three decades and are clearly targeted at an economy increasingly driven by AI models, software and digital assets rather than purely physical technologies.
Under the updated framework, “technical information” eligible for trade secret protection now expressly covers data, algorithms, computer programs and code alongside traditional categories such as formulas, processes, materials and technical methods. This shift reflects the reality that competitive advantage in sectors such as artificial intelligence, cloud computing, fintech and advanced manufacturing often resides in proprietary datasets, model architectures and optimization techniques that are never publicly disclosed.
At the same time, the core definition of a trade secret remains broadly consistent with China’s Anti Unfair Competition Law. To be protectable, information must:
Not be publicly known or readily obtainable by relevant market participants
Have commercial value
Be subject to reasonable confidentiality measures.
The regulations clarify what “reasonable” measures can look like in modern digital environments, including access controls, encryption, data classification, confidentiality agreements, internal training and restrictions on copying or downloading sensitive information. They also address remote work and cross border collaboration by recommending tiered permissions, data masking and audit logging for digital systems.
For AI developers and data driven businesses, the updated rules underscore the importance of robust trade secret governance. Unlike patents, which require public disclosure and may not easily capture iterative improvements to models, trade secret protection allows companies to safeguard non public elements of their technology stack—such as training datasets, feature engineering, parameter tuning strategies and deployment pipelines—provided confidentiality is actively maintained.
From an enforcement perspective, the regulations expand the catalogue of “improper means” to specifically cover electronic intrusion, remote data scraping, unauthorized access to digital office systems and exploitation of software vulnerabilities to obtain trade secrets. They also strengthen “full chain” liability by introducing provisions on instigation, inducement and assistance in trade secret infringement, and extend administrative enforcement to certain overseas acts that disrupt China’s domestic market order. Together, these changes are intended to give regulators sharper tools to respond to technology leaks, particularly in strategic sectors where AI and data assets are viewed as core national and commercial interests.
Source:
https://www.chinadaily.com.cn/a/202606/04/WS6a20d161a310d6866eb4c586.html
(English)