14 September 2026 , 19:55
420
0
China’s spy agency has warned that AI could pose a risk to the country’s political and social security, signalling growing concern in Beijing that rapid advances in the technology could be exploited against the country.
In its first public statement on AI, the Ministry of State Security on Sunday detailed the threats the technology presented, including to China’s cyber defences, sensitive data and military apparatus.
“Hostile forces and people with ulterior motives are abusing generative AI technologies such as deepfake audio and video, AI-generated text and images, and ‘intelligent troll armies’ to cheaply and on a mass scale fabricate political rumours, disseminate harmful information and incite divisive sentiments,” said state security minister Chen Yixin.
“These activities constitute an attack on public opinion and cognitive warfare against China, directly threatening our political security, institutional security and ideological security,” he added.
The statement marks a significant expansion of Beijing’s public discussion of AI safety. Typically, the debate has been led by the main regulator, the Cyberspace Administration of China, and focused on regulatory and technical concerns.
The statement comes as US technology leaders have voiced alarm over the pace of frontier AI development, following increasingly stark warnings from researchersabout the possibility of catastrophic outcomes, including an extinction event.
The MSS statement did not contain warnings that AI could cause human extinction and did not call on Chinese companies to slow their work in the area.
Over the weekend, Anthropic’s boss Dario Amodei penned an essay calling for the tech sector to “slow the pace at which we improve the capabilities of AI models”. Sam Altman and Elon Musk both issued statements supporting the proposal.
The Communist Party nationalist tabloid Global Times on Monday attacked the suggestion in Amodei’s essay that a “Chinese lead in AI would pose grave danger for the United States and the world”.
Amodei’s claim “risked turning AI safety into a geopolitical zero-sum issue rather than addressing a global challenge that requires co-operation among major AI powers”, the paper said, adding that Amodei’s motivation was “commercial interests”.
Several regulatory officials, diplomats and academics who work on tech policy and AI governance say that Beijing has grown increasingly concerned about its cyber vulnerabilities, given the advances made by Anthropic and OpenAI, whose new models have demonstrated advanced hacking capabilities.
Chinese officials have also raised AI safety concerns in discussions with US counterparts and companies in recent weeks, the diplomats and academics say.
In these talks, officials have cautioned against any move by Washington to ban open-source Chinese models. Presidents Donald Trump and Xi Jinping are set to meet later this month in the US, where AI safety is expected to feature high on their summit’s agenda.
In Sunday’s statement, Chen also outlined the risk of Chinese companies leaking “sensitive information” through their use of foreign AI tools. That warning could presage tighter restrictions on the use of foreign AI models in sensitive sectors.
Chinese state-owned enterprises and companies operating in areas such as finance and healthcare already face restrictions on the use of foreign technology, although enforcement varies.
Head of Investigations