The Chinese Communist Party’s flagship theoretical journal, Qiushi, has published the full text of a speech by President Xi Jinping that marks a significant shift in the leadership’s rhetoric regarding artificial intelligence (the full speech in Chinese can be found here). Delivered earlier this year, the speech explicitly identifies the “loss of control over technology” as a primary risk associated with the rapid development of future industries. This phrasing marks the first time China’s top leader has publicly articulated the existential anxieties often associated with advanced AI systems, aligning Beijing’s internal discourse more closely with the global AI safety conversation than at any previous time.
The speech, titled “Forward-looking planning and development of future industries,” was originally delivered on January 30, 2026, during the 24th collective study session of the 20th Central Committee Politburo. Its publication in Qiushi on May 31 elevates the text to the level of core ideological guidance for the entire party apparatus, meaning that officials at every level of government and industry are now expected to internalize and act upon its directives. According to the China AI Bulletin, the document outlines six “future industries” targeted for the upcoming 15th Five-Year Plan: quantum technology, biomanufacturing, hydrogen and nuclear fusion energy, brain-computer interfaces, embodied intelligence, and 6G.
A Shift in Risk Framing
The most notable element of the speech is found in its fifth point, where Xi addresses the necessity of balancing development with security. He calls on officials to “coordinate development and security, explore scientific and effective methods of regulation, systems for technology monitoring, risk early warning, and emergency response.” Crucially, he urges the party to “anticipate and respond to new types of risk such as loss of control over technology (技术失控), ethical breaches, and data abuse.”
The use of the term “loss of control” (技术失控) is a stark departure from previous official statements. Historically, Xi and other senior leaders have framed AI risks in terms of safety, security, reliability, and controllability—concepts that imply the risks are manageable through proper engineering and governance. The phrase 技术失控 carries a more alarming connotation, suggesting a scenario in which technology develops beyond human capacity to oversee or correct. It is worth noting that Xi did not explicitly tie 技术失控 to AI specifically in the speech text; the term appears in a broader governance discussion covering all six future industries. Nevertheless, Xinhua follow-up coverage linked the phrase directly to AI, and the term has appeared in AI-specific regulatory documents including TC260’s AI Safety Governance Framework v2 (September 2025). A July 2025 report by Concordia AI on the state of AI safety in China specifically noted that leadership statements had “not explicitly mention[ed] risks from misuse of advanced AI systems or from loss of control of superintelligent AI.” The inclusion of this specific phrase in a Qiushi publication indicates that concerns about runaway AI have reached the highest levels of the Chinese government.
The “Blind Men Touching an Elephant” Analogy
To underscore the complexity and unpredictability of these emerging technologies, Xi employed a classic Chinese idiom. He warned against the approach of “blind men touching an elephant” (盲人摸象) when dealing with scientific and technological change. This analogy suggests a concern that regulators and developers may only grasp isolated fragments of the AI landscape, failing to comprehend the full scope and potential consequences of the technology they are building. It is a rare moment of epistemic humility in official Chinese discourse on technology, which has more typically emphasized confidence in the state’s capacity to guide and control technological development.
This rhetorical framing suggests that Beijing recognizes the limitations of its current regulatory toolkit when faced with the rapid iteration of generative models and embodied intelligence.
(Related: China Closed the AI Gap With America While Running the World’s Strictest AI Rulebook)
The call for “systems for technology monitoring” and “risk early warning” implies that the government intends to build more sophisticated, proactive mechanisms for overseeing AI development, rather than relying solely on reactive censorship and compliance checks. This could translate into new regulatory requirements for AI developers, including mandatory reporting of capability thresholds, red-teaming exercises, and incident disclosure obligations.
Implications for China’s AI Sector and Global Governance
The publication of this speech serves as a clear directive to both regulators and the domestic technology industry. While the government remains fully committed to dominating the six named “future industries,” including embodied intelligence, this ambition will be tightly coupled with stringent security mandates. Companies developing advanced models will likely face increased scrutiny regarding the alignment and controllability of their systems, and the speech provides the ideological justification for regulators to impose new requirements that might otherwise be seen as obstacles to innovation.
The explicit mention of “loss of control” also aligns China’s internal discourse more closely with the global conversation on AI safety, which has increasingly focused on existential risks from advanced AI systems. (Related: Trump Signs AI Executive Order Requiring Pre-Release Review) As the United States and China begin formal intergovernmental dialogues on AI governance, Xi’s acknowledgment of these profound risks provides a shared, albeit tense, foundation for those discussions. If both governments genuinely believe that AI could develop in ways that escape human control, there is at least a theoretical basis for cooperation on the most severe risk scenarios, even amid fierce competition in commercial and military applications.