As artificial intelligence integration accelerates across China’s digital economy, a core technical component of large AI models has quickly moved from behind-the-scenes infrastructure to the center of a national security alert, as authorities warn of widespread fraud, data theft and systemic risks tied to the unregulated handling of AI tokens, known locally as ciyuan.
Defined as the smallest unit of data processed by large generative AI models, tokens serve multiple critical functions in modern digital ecosystems. Beyond measuring computing power usage for AI services ranging from text generation to image and video editing, tokens also act as digital credentials for identity verification, access control to protected platforms and payment authorization for AI API calls. Their rapid proliferation has mirrored the boom of China’s generative AI sector, with official data showing average daily token call volumes surpassed 140 trillion as of March 2026 — a more than 1,000-fold increase from the start of 2024.
This explosive growth caught the attention of China’s Ministry of State Security, which released a public warning on April 8 detailing three major categories of token-related threats facing individual users and national infrastructure alike: token theft and hijacking, unauthorized forgery and tampering, and large-scale fraudulent investment schemes.
Unencrypted tokens, the ministry explained, are highly vulnerable to exploitation through common cyber threats including malware injections, coordinated network attacks and connections over insecure public networks. Once stolen, tokens allow criminals to impersonate legitimate users, gain illegal access to sensitive personal and institutional data, and carry out unauthorized financial transactions without detection. For platforms with insufficient verification protocols, attackers can also easily forge or alter token credentials to bypass built-in security checks, gaining entry to restricted systems.
Fraudulent schemes have emerged as one of the most pervasive threats to the general public, as scammers capitalize on low public awareness of what tokens actually are. On popular social media platforms, bad actors have been promoting low-cost token packages, unlimited usage plans and unauthorized token agency resale schemes, marketing tokens as a high-profit investment opportunity similar to the hyped virtual currency schemes of previous years. Many of these promotions use standardized, persuasive scripts to frame the emerging “token economy” as a quick path to wealth, when the operations are actually pyramid schemes designed to recruit new participants rather than generate legitimate returns.
Huang Daoli, a senior researcher at the Third Research Institute of the Ministry of Public Security, noted that what was once a niche technical term confined to AI development circles has now become a foundational building block of China’s digital economy. “In practice, tokens are used to measure generative AI computing power and bill API calls,” Huang explained. “In many applications, they also serve as credentials for identity verification and access control, giving them simultaneous technical, transactional and security functions.”
With daily usage already crossing the 140 trillion threshold, Huang emphasized that token security risks can no longer be dismissed as isolated, minor technical issues. “If token security is breached at scale, the impact may spill over from personal privacy violations and individual financial loss to broader threats to national data security and even overall economic security,” she warned.
Scams that repackage tokens as legitimate investment products exploit the public’s limited technical understanding, Huang added, and pose direct risks to personal privacy and identity security. Widespread token forgery, meanwhile, could undermine data governance protocols in critical sectors including government administration and financial services, creating systemic risks for the country’s digital infrastructure.
China already has a robust legal framework in place to govern digital security, including existing cybersecurity and data protection laws that cover token-related activities. According to Huang, the top priority moving forward is strengthening enforcement of these rules, particularly through tighter identity management and increased oversight of unregulated high-risk API resale markets.
Both the Ministry of State Security and industry experts have issued clear guidance for the public to avoid token-related risks. Authorities stress that tokens function exclusively as digital credentials for AI services, not as investable financial products. Users are advised to steer clear of get-rich-quick token schemes, only access AI services through verified trusted platforms, and strictly protect token credentials and account passwords just as they would for online banking tools.
Huang emphasized that tokens should be classified as highly sensitive digital assets on par with critical payment tools. “The key is to change the mistaken belief that an API key is merely a trivial technical parameter,” she said. “It should be treated as what it is: a critical data asset that requires rigorous protection.”