China's Ministry of State Security, once shrouded in secrecy, is now adopting a more communicative stance to warn against the dangers posed by artificial intelligence. The spy agency identifies AI as a systemic vulnerability capable of facilitating confidential data leaks or large-scale opinion manipulation through social engineering. This distrust highlights a glaring paradox: while China asserts itself as a global leader in the sector, it fears these same tools could become vectors of influence for foreign powers. The challenge for Beijing is to maintain its dominant position while imposing an absolute security lockdown on usage and information flows.
To address these threats, authorities have significantly tightened the national legislative framework. The 2023 revision of the anti-espionage law now extends the definition of state secrets to a vast spectrum of data, potentially encompassing the databases used to train models. This is compounded by strict regulations on the labeling of AI-generated content, effective since September 2025. This strategy directly aligns with the doctrine of Xi Jinping, which demands technology that is "safe, reliable, and controllable." The industrial breakthrough of models like DeepSeek, which recently shook global financial markets by wiping out hundreds of billions of dollars in market capitalization, illustrates the technological prowess the regime seeks to domesticate.
This repressive approach is not unprecedented and mirrors the trajectory imposed on the digital asset sector. Since the systemic ban on cryptocurrency trading in September 2021, Chinese authorities have sought to replace decentralized networks with state-supervised infrastructure. The digital yuan, whose international operations center was inaugurated in Shanghai in 2025, embodies this desire for total financial traceability. In parallel, the regulation of stablecoin issuers in Hong Kong demonstrates a desire to capture innovation while blocking the path for autonomous protocols and self-hosted wallets that evade central oversight.
The financial stakes linked to this digital transition remain colossal and sometimes opaque. Despite the restrictions, China remains a major player in the crypto landscape, theoretically holding approximately 194,000 BTC seized during the PlusToken case, although the exact fate of this digital treasure is regularly questioned by on-chain analysts. Controlling value on open networks becomes even more complex as decentralized computing infrastructures, such as Bittensor or Render, emerge. For the Chinese state, these protocols represent a risk of circumvention similar to open-source AI models, which can be redistributed outside any jurisdiction the moment they are published.
Ultimately, China finds itself at a crossroads between the ambition for technical supremacy and the imperative of domestic stability. The increased surveillance of language models and the ban on public crypto networks share a single goal: to prevent the emergence of decentralized systems capable of breaking free from state supervision. The dilemma is profound, as technological innovation generally thrives on openness. By attempting to place AI under strict tutelage, as it did with Bitcoin, China risks stifling its own capacity for innovation in favor of a rigid and centralized digital sovereignty.