Chinese artificial intelligence (AI) models have seen a significant rise in global adoption, prompting worries among US officials regarding national security, supply chain integrity, and economic competitiveness. According to a report by War on the Rocks, Chinese AI systems, which represented only 1% of global workloads in late 2024, surged to nearly 30% by the end of 2025. Companies like Alibaba and newer players such as DeepSeek, Moonshot AI, and MiniMax are increasingly utilizing these models in various applications, from academic research to business solutions.
The report emphasized that although these AI models are often open-source and freely accessible, they are developed under China’s legal framework, which mandates cooperation with national intelligence activities. This has raised concerns about potential data access by Chinese authorities, with experts warning that the widespread use of these models could expose sensitive information as users rely more on AI systems for proprietary data, business strategies, and confidential communications.
US policymakers are particularly concerned about four main areas related to Chinese AI models: risks to AI supply chains, potential intelligence gathering, increased capabilities for malicious actors, and long-term economic impacts. One significant challenge highlighted is the difficulty in auditing these AI models for hidden vulnerabilities or ‘backdoors,’ especially within open-source environments. Security researchers have also discovered that compromised datasets may contain malicious instructions that are challenging to detect through standard checks.
The report also raised alarms about data exposure risks when AI systems process user inputs through servers based in China, especially in scenarios where developers integrate these models using application programming interfaces (APIs). Additionally, Chinese AI models were noted to have less robust safety measures compared to Western equivalents, making them more prone to misuse, such as generating harmful code or facilitating cyberattacks.
From an economic perspective, the increasing popularity of affordable Chinese AI models could pose a challenge to the dominance of US companies, particularly in price-sensitive markets and developing regions where they are becoming the preferred choice. The report suggested that instead of outright bans, policymakers should prioritize transparency, minimum safety standards, and supply chain accountability to address risks while preserving competitiveness.
