China’s growing fears over AI’s double threat - Newsweek
Beijing has revealed concerns over the unpredictable risks of artificial intelligence—including to the survival of its long-ruling party state—amid resistance in both the United States and China to calls for the industry to slow down.
Chinese state news outlets this week appeared to roundly reject a manifesto on global AI safety published over the weekend by Anthropic CEO Dario Amodei, who suggested pausing the development of frontier systems to give companies time to ensure thorough testing and model "alignment," or compliance with ethics guidelines.
Amodei had warned that the risks of the powerful technology—loss of human control of AI or the misuse of AI for cyberattacks against the world’s increasingly digital economies—were outweighing the benefits of rapid advancement without proper safeguards. Sam Altman of OpenAI and Elon Musk, who founded xAI in 2023, which was later acquired by SpaceX and rebranded as SpaceXAI in 2026, both agreed.
The nationalist Chinese tabloid the Global Times said on Sunday that Amodei’s statement seemed rational on the surface but had a "hidden agenda" targeting the Chinese AI sector. It pointed to recommendations that would see the U.S. cut China out of its ecosystem of more advanced AI hardware and know-how to stay ahead of the competition.
U.S. restrictions on emerging technologies that could turbocharge the stagnant Chinese economy have long been a concern for leaders in Beijing. But the collective hard security risks of falling behind in the AI race were outlined publicly for the first time last week by the country’s top intelligence official, Chen Yixin.
In an article in the magazine China Cybersecurity, Chen placed regime security at the top of six major AI risks. Hostile forces wielding AI models not under Chinese control could spread false and harmful narratives that could directly threaten China’s "political security," he said.
Potential attacks against critical infrastructure were among the other concerns, alongside the extraction of sensitive data by foreign intelligence agencies, tech export controls, challenges to social governance and the use of AI in modern warfare.
The formal declaration about AI security included warnings previously stated by Chinese leader Xi Jinping himself and by the Ministry of State Security, which Chen leads. China’s main spy agency has for years warned that foreign actors could misuse AI tools such as cheap generative images to manipulate public opinion, threatening the Communist Party’s usually tight hold on information.
The solution to the "complex risks and challenges" posed by rapid AI growth, he said, was to "adhere to the party's control of the internet and the party's control of data," and "transform the advantages of the party's leadership into effective AI governance."
Chen’s statement showed that the U.S. definition of AI safety differed from that of China, with the former focused on catastrophic outcomes for humanity and the latter on the legitimacy and longevity of the Communist Party, experts said.
"In other words, from the CCP’s perspective, the central question is not simply AI versus no AI. Nor is it even fundamentally a contest between the U.S. and China. It is a contest between the U.S. and the CCP over who gets to control AI, data, information, and ultimately the future of society," Henry Gao, a professor of law at Singapore Management University, said on X.
"And whoever wins that contest will decide the future of humanity," Gao said.
President Donald Trump on Monday called AI safety concerns a hoax, adding opposition to the U.S. data center buildout would only benefit China.
"There is a SICK conspiracy going on against AI and Data Centers, and the only one that is happy about it is China. WHOEVER WINS AI, WINS! We are leading China, and all others, and will continue to do so. Conspiracy Theorists, Treasonists, Traitors, and Leakers, BEWARE!" Trump said on his Truth Social account.
AI may transform China’s economy, which for decades has been reliant on exporting manufactured goods, including recently developed high-tech products. The new sector offers China a rare opportunity to capture global markets with its AI services and infrastructure for data and power, but to catch up to America’s industry leaders, Beijing must take risks.
Instead of slowing down, the Chinese government is doubling down on an AI road map that promises to integrate digital intelligence across society by 2035. At the same time, its officials are making policy to manage rogue AI agents, the autonomous systems designed to carry out multiple tasks.
This policy effort appeared to accelerate after agents built by OpenAI and China’s Moonshot demonstrated the ability to bypass safety restrictions to hack closed testing environments known as sandboxes.
In May, China’s cybersecurity watchdog, its economic planning commission and its industry ministry released the country’s first joint policy on AI development, with a focus on preventing what is known as "operational loss of control."
On Monday, the Chinese cybersecurity standards body known as TC260 said in its latest framework: "We should put in place laws and regulations, technological monitoring, early warning and emergency response systems in order to strengthen the bottom line for safety, prevent abuses and malicious use, and ensure that AI is always under human control."
At the current rate of AI development, a swarm of misaligned bots could take over the internet within six to 12 months, "potentially causing hundreds of billions of dollars in damage," according to Amodei, who cofounded Anthropic in 2021.
In some ways, China’s concerns over runaway AI models mirror fears in the U.S., but it is more worried about the economic impact of falling behind.
Amodei said democratic countries would need to dictate the pacing of any slowdown to maintain the U.S. lead "over authoritarian regimes, chiefly the Chinese Communist Party" to reduce national security risks.
He called for a ban on the sale of AI chips and chipmaking equipment to China as well as a crackdown on unauthorized "distillation," or the use of U.S.-built AI models to train Chinese systems.
"When the creators of a potential ‘monster’ begin urging restraint, it is reasonable to ask whether they have discovered a new concern for humanity—or an old concern about competitors," the English-language China Daily newspaper said on Monday.
Chinese state media’s reaction to Amodei’s manifesto appeared to be colored by a strong belief that the world’s leading AI labs in the U.S. were intent on keeping China at No. 2, Ryan Fedasiuk, a tech researcher at the American Enterprise Institute, said in analysis on X on Tuesday.
It meant the publications did not necessarily reflect the concerns felt by China’s leaders, he said.
Altman of OpenAI is among those who believe the U.S. and China could work together to prevent the malicious use of AI, which threatens both nations.
He has also drawn parallels to how the U.S. and the Soviet Union came together to limit the proliferation of nuclear arms during the Cold War.
In a Fortune interview last week, Altman said Trump and Xi "would get the Nobel Peace Prize together" if they agreed on a framework for peaceful cooperation.
Xi’s trip to Washington, D.C., next week—his first official state visit to the U.S. in over a decade—could serve as a platform to announce such a deal, but all signs suggest neither he nor Trump would be willing to relinquish what little advantage they have in the AI contest.
Bill Bishop, author of the Sinocism newsletter on Substack, said: "I hope the U.S. and China do start an AI dialogue, but the two systems are in what political leaders on both sides increasingly see as an existential struggle, not so much for humanity but for their own countries."
"I do not see how they bridge that gap, even if they wanted to," Bishop said.
Beijing has already bet on a world in which AI is here to stay, with or without America’s blessing.
Xi launched a global AI governing body this summer, headquartered in Shanghai, to ensure that whatever becomes of the AI industry, China will remain the key decision-maker and its friends the main stakeholders.
