Anxiety about AI in the United States appears to be getting worse. Prosaic concerns about machines disrupting industries, stealing jobs, and compromising schoolwork are giving way to more harrowing fears about the end of humanity.

Even after top AI executives in the U.S. called for slowing down the breakneck pace of development, President Trump dismissed these safety concerns as a “hoax” that would hand China an advantage in the race for AI supremacy.

Outside of China, Trump’s assertion may appear well founded. While the U.S. roils over the prospects of AI, China’s leadership looks to be charging ahead. In surveys, Chinese people appear more enthusiastic about AI and its potential benefits than respondents in the U.S., and they are adopting it more quickly. Although Chinese President Xi Jinping has expressed more alarm than Trump about the threats of AI—and has even called for regulation to “ensure that AI is always under human control”—his government has made adopting the technology a national priority. The plan is to have AI in every aspect of China’s economy and society by 2035. The country’s steady release of AI models that are open-source, high-performing, and relatively cheap to build and run has been sending shivers through Silicon Valley and Wall Street.

Read: What really happens if China wins the AI race?

But for all the concern about China’s advances in AI, the reality is that the country is racked by the same anxieties about the onrush of AI as the rest of the world. Chinese social-media platforms are filled with posts in which workers and professionals express their worries that AI will steal their job and livelihood. Leaders in Beijing seem ever more concerned about the potential political and economic risks of this technological disruption.

In the U.S., these tensions are being debated openly and, in theory, will be managed in the courts and on Capitol Hill. In China, where there are no such avenues for people to determine their own fate, leaders must still confront the problems and threats inherent to this technological revolution. This puts the Communist Party in a bind: Advancing AI may well empower China on the world stage, but holding the technology back may be necessary for the party to preserve itself.

Xu Peng, a 30-year-old actor, was out of work last year when he got an unexpected break. A director had reached out to cast him in a short-drama series, each episode minutes long, which have become enormously popular in China. Xu traveled to Hengdian, a film-production hub in eastern China, and began landing roles in one or two dramas a month. And then it all stopped. Xu had expected to act in two dramas after the Chinese New Year in February, but never heard from the studio. When he contacted the producer, he got the news: The company had canceled the productions and would be making videos using artificial intelligence instead.

“It was as though everything was fine yesterday, and today it is suddenly gone,” Xu said.

Xu’s fate is the fear of workers across the country—and around the world. Social-media posters regularly express anxiety about being “distilled,” a term Chinese people use for when an employer uses a worker’s data to train an AI model. In May on Douyin, the Chinese version of TikTok, someone going by the name Nanbei wrote that after leaving a recent job, she learned that her former employer had fed thousands of her chat messages, hundreds of documents, and every one of her work emails into an AI system to create a “digital robot” that took over her responsibilities—and even gave it her name. “The company no longer has to pay me a cent in salary, yet it can keep me there ‘working overtime’ forever,” Nanbei wrote. Another worker, using the handle Xiexiaoya, posted on RedNote, China’s Instagram, in June that the prospect of being distilled kept her awake at night. “I’m afraid my AI double might become more popular than I am,” she wrote.

Two years ago, Shen Anyu, a voice actor who earns a living from narrating news commentaries and other short programs, began noticing an alarming number of videos on social-media platforms that seemed to feature his voice without his permission. Shen hired a lawyer earlier this year and filed a lawsuit against a voice-generation platform. The case is still pending, but in the meantime, his career is “losing blood,” he said, as he loses work to AI. Living with the new technology “is inevitable,” he said, but “I’m not yet sure how we should coexist with AI.”

Though some Chinese people are embracing AI, many are not ready for the change it brings. Bruce Jiang, a 42-year-old commercial director in Shanghai, loves choosing the right lens and finding the perfect shot, and he didn’t want to outsource these and other aspects of his craft to an algorithm. But after his clients began pressuring him to use the technology—“‘You only have to tell AI what you want, and it produces it like magic,’” he said they insisted—he began using tools such as the video-generator Kling for certain tasks. The experience, he said, has been demoralizing. The old method “was work, and it was a cost, but it was also a kind of pleasure,” Jiang said. “That pleasure is disappearing.”

Chinese policy makers at national and local levels have introduced ambitious plans and financial support to rush the country into the AI age. Entrepreneurs across China are adopting AI in the hopes of cutting costs, remaining competitive, and aligning their firms with Beijing’s priorities. In China’s cutthroat society—and especially in sectors where AI is making rapid inroads, such as the entertainment industry—many like Jiang are afraid that they’ll get left behind if they don’t join in.

But these swift and radical changes have yielded mixed results. The state-linked China Center for Information Industry Development estimated in June that nearly 70 percent of the country’s small businesses have not moved beyond initial trials of adopting AI in their operations, largely owing to the costs and complexity of integrating the technology.

After the Chinese AI lab DeepSeek released a large language model last year, Chinese companies rushed to adopt it. But instead of buying a software subscription (an unpopular practice in China owing to the country’s history of piracy), some businesses opted for what locals call an “all-in-one machine,” or a mini data center with DeepSeek installed, which managers simply plugged into their computer systems. This solution was especially popular among state-owned enterprises and government bureaus looking to appear advanced, but the hardware proved to be an ineffective way to use AI. Buyers soon came to see that the machines weren’t worth the investment, and their sales quickly fell.

Earlier this year, plenty of buzz similarly surrounded OpenClaw, a free, open-source, autonomous AI agent. Tech companies hosted popular events to help users install and use the tool, a process that Chinese enthusiasts call “raising a lobster.” At the OpenClaw festival in Beijing in March, the internet-service provider Baidu set up pools with live crustaceans to entertain attendees. Local-government officials encouraged the frenzy by pledging financial aid to start-ups that used OpenClaw.

Disillusionment came quickly. Some distraught users took to social media to express their alarm at how the AI agent, which required masses of personal data to operate without detailed instructions, had a habit of going rogue. “I jumped on the bandwagon” expecting to get a “digital employee,” a woman using the social-media handle HotdogWang wrote in March on RedNote. “Instead, it turned out to be both a money pit and a ticking time bomb.” Some reported that their AI helper had deleted emails or made credit-card purchases without permission. One Douyin user, posting under the name Qizaibushidaju, wrote, “I simply installed it because everyone else was doing so,” only to uninstall it 72 hours later owing to security concerns.

Although Chinese tech executives are not yet engaged in a public debate over the risks of an algorithm-created calamity—the country’s constraints on free speech may inhibit them from speaking out—some Chinese people do worry about more than just jobs and data. “There are only two options for the future,” a man calling himself Qiangni Yinshou wrote on Douyin. “Either humanity goes extinct, or humans shut AI down.”

A user by the name of Buyehou wrote on RedNote that when he engaged a chatbot in a philosophical discussion about humanity’s future, he was told that the options included “either merging with AI or being eliminated.” The man added that “when AI becomes far smarter than humans, has consciousness of its own, and can evolve independently, it will be impossible for humans to control it. There is no turning back now.”

Despite their outward show of confidence, China’s leaders plainly don’t know how life works with AI either. But figuring it out is an urgent concern here, given Beijing’s evident faith in AI as a magical solution to the country’s deepening economic and social problems, and as an essential tool in the pursuit of global dominance over the United States. In the U.S., the diffusion of AI is largely a bottom-up phenomenon as companies, investors, and consumers test the risks and benefits of this new technology. In China, where the development and spread of AI is a national mission that is costing the government hundreds of billions of dollars, leaders are eager to impose these advances from above.

In theory, China’s autocrats can ram through their agenda with scant regard for public sentiment. If confronted with the kinds of protests against AI data centers that have become prevalent in the U.S., Beijing’s leaders could easily send police to crush them and then cover up the problem in a deluge of propaganda. The Communist Party has already shown its willingness to override public concerns in pursuit of technological advantage. In 2024, trials of autonomously driven taxis in the city of Wuhan sparked concern over how robotaxis were “stealing people’s rice bowls”—but the trials have expanded anyway.

Yet China’s leaders have traditionally prized policies that promote stable employment out of fear that mounting joblessness could create unrest and threaten the political order. When the country first embraced markets and opened to trade, its rapidly growing economy and expanding private sector could absorb displaced workers. Now, when China’s sluggish economy already has a jobs problem—nearly a fifth of young workers are unemployed—the advances of AI threaten to put many more people out of work.

Read: The race for global domination in AI

“In a nominally Communist state, the image of the party elite allying with business interests to displace the human proletariat with a robot class of workers carries certain political risks,” Carl Minzner, a senior fellow at the Council on Foreign Relations, told me. “There is the potential for a lot of disgruntled people.”

China’s leaders are showing signs of nervousness. Earlier this year a court in Hangzhou, a technology hub in eastern China, upheld a ruling in favor of a tech worker who had challenged his compensation package after being replaced by AI. In the court’s judgment, the dismissal was unlawful and the worker deserved more money. Although a local-court verdict does not set a national precedent, it appears to be a warning to employers that mass AI layoffs are politically unacceptable. “As employers enjoy the benefits of technology, they must also shoulder corresponding social responsibilities,” the Beijing municipal government wrote last year. The state news agency Xinhua commented in March that “those companies that equate AI with ‘reducing staff’ may seem to lower costs in the short term, but in reality, they lose the core competitiveness of talent accumulation and further erode employee trust.”

Pushing companies to invest in new technologies without shedding redundant workers may preserve social order, but it would also limit AI’s benefits. Jeffrey Ding, an expert on AI in China at George Washington University, told me that the country’s leaders face a “balancing act” between political and economic necessity. “AI does have the potential for bringing massive, widespread productivity growth,” he said. But he also noted that the Communist Party prioritizes legitimacy and regime stability. If embracing AI demands “massive displacement and a lot of structural changes, then that’s where the trade-off begins to bite, and they might have to veer toward slowing down AI adoption.”

If American AI firms do, in fact, slow the pace of innovation, the Communist Party may see this as an opportunity to race ahead in AI. But China’s leaders, facing their own domestic challenges and preoccupied with their own grip on power, may decide that imposing these innovations on the masses is simply too risky—for their own sake, if not for the sake of humanity.

That would probably suit many Chinese people just fine. Xu, the actor, has found work teaching archery and horseback riding—skills he acquired in acting gigs. But he hasn’t given up on his theatrical career. He now plans to produce his own short dramas, partly with the help of AI, though he said that the performers on-screen will all be living, breathing people. AI “will not completely replace human actors,” Xu insisted. At least he can hope.

Cao Li in Hong Kong contributed reporting to this story.