Data, algorithms added to trade secret regulations

China brought into force on Monday new trade secret regulations that explicitly extend legal protection to data and algorithms for the first time, as the country moves to strengthen safeguards for digital assets amid intensifying competition in artificial intelligence and advanced technologies.

The new rules, issued by the State Administration for Market Regulation, the country's top market regulator, marked one of China's most significant updates to its trade secret regime in years, reflecting the country's effort to adapt intellectual property protections to an economy increasingly driven by data, AI models and digital innovation.

Under the new regulations, technical information eligible for protection now explicitly includes data, algorithms, computer programs and code, alongside more traditional categories such as formulas, manufacturing processes, materials and technical methods.

The move comes as Chinese companies invest heavily in artificial intelligence, cloud computing and digital infrastructure, sectors where competitive advantages increasingly depend on proprietary algorithms and data resources rather than physical assets alone.

Wang Peng, a researcher at the Beijing Academy of Social Sciences, said:"For the first time, digital elements such as data, algorithms, computer programs and code have been formally incorporated into China's trade secret protection framework.

"This provides a clearer institutional basis for protecting core technologies and digital assets in the digital era."

The update represents a notable shift from the traditional focus of trade secret protection, which historically centered on industrial know-how such as manufacturing techniques, engineering drawings and product formulas.

As the digital economy becomes a major engine of growth, however, companies are increasingly deriving their competitive edge from accumulated data resources, optimized algorithms and digital operating capabilities.

Thus, the new regulations make it clear that not all data and algorithms automatically qualify as trade secrets.

To receive legal protection, the information must meet three conditions: it must not be publicly known, it must have commercial value and the owner must have adopted reasonable confidentiality measures.

The rules define "not publicly known" as information that is generally unknown and not readily obtainable by relevant industry participants at the time an alleged infringement occurs.

The regulations arrive as competition in generative AI accelerates globally. Chinese firms including DeepSeek and Alibaba have stepped up efforts to improve model performance and efficiency, while policymakers increasingly emphasize technological self-reliance and innovation.

Unlike patents, which require public disclosure and may not fully protect iterative model development, trade secret protection allows companies to safeguard proprietary model architectures, training methods, optimization techniques, and other commercially valuable innovations.

For artificial intelligence developers, internet platforms, fintech companies and advanced manufacturers, algorithms often embody years of investment in talent, computing power and research.

"The new framework is also expected to provide companies with stronger legal tools to protect core intellectual assets while encouraging continued investment in research and development," Wang said.

"By formally recognizing data and algorithms as potential trade secrets, China is signaling that in the AI era, digital assets may deserve the same level of protection once reserved primarily for factories, patents and industrial know-how," he added.

chengyu@chinadaily.com.cn