SELECT LANGUAGE BELOW

China’s AI is genuinely artificial – and the US needs to compete against Xi’s zero-sum tech race and appropriated progress.

China's AI is genuinely artificial – and the US needs to compete against Xi's zero-sum tech race and appropriated progress.

Even with all the excitement about China’s advancements in artificial intelligence, much of their progress appears to rely on theft and smuggling. The Chinese Communist Party (CCP) is firmly focused on keeping everything under its control. This situation is becoming more apparent as this week’s summit in Beijing approaches, with leaders Donald Trump and Xi Jinping set to meet.

Xi seems eager to position China as a frontrunner in so-called “breakthrough technology.” He may be feeling pretty confident since Trump is tied up with the conflict involving Iran and has limited ways to respond to China’s increasingly bold maneuvers.

Just last month, U.S. officials accused China of engaging in “industrial scale” theft of knowledge from American AI labs. Additionally, U.S. prosecutors say they’ve dismantled an international smuggling operation that sent advanced chips worth billions to China, disregarding sanctions.

The CCP has also intensified its efforts to protect homegrown AI innovations by blocking Meta’s attempt to acquire the Chinese AI startup Manas for $2 billion. Naturally, the founders of Manas were stopped from leaving the country.

cheaper and faster

The allegations of theft center around a method known as “distillation.” Essentially, China is alleged to be training smaller AI models unlawfully based on larger, costly U.S. models.

An internal memo leaked indicated that the “U.S. government has received information showing that foreign companies, mainly from China, are engaged in a large-scale effort to distill America’s advanced AI systems,” as noted by Michael Kratsios, who leads the White House Office of Science and Technology Policy.

This distillation process involves setting up numerous fake accounts for a specific AI tool to gather information collectively.

Anthropic, a U.S. AI firm, reported that its powerful Claude chatbot has engaged in over 16 million interactions and identified 24,000 fraudulent accounts. The paper claimed that leading Chinese research institutions were behind efforts to acquire significant capabilities “in a fraction of the time and at a fraction of the cost.”

Moreover, the company highlighted potential risks, noting that “refined” applications lack the necessary safeguards against using AI for harmful activities, including the creation of biological weapons and large-scale cyberattacks.

It appears the Chinese government has established an extensive, well-funded smuggling operation to evade U.S. restrictions on top-tier NVIDIA chips needed for AI model training.

Recent federal indictments against Chinese nationals describe how servers loaded with “billions of dollars” of restricted chips were shipped to powerful companies in Southeast Asia, then repackaged and sent to Hong Kong and mainland China.

One surveillance clip showed a defendant using a hair dryer to switch out adhesive labels and serial number tags, a surprisingly low-tech image in contrast to the high-stakes smuggling operations supposedly endorsed by Xi, who promotes a “race to the top.”

“Singapore Washing”

The ongoing charges are viewed as just the beginning of uncovering a larger smuggling network, and the battle for AI dominance is framed as a significant factor that could redefine global power dynamics.

Xi is not just set on ensuring China comes out on top; he is also keen on keeping AI firmly under the CCP’s authority, as evidenced by the fallout for Mark Zuckerberg.

Meta executives were under the impression that acquiring Manas was a done deal, with the startup’s team already transitioning to Meta’s office in Singapore. Manas functions as an AI agent, performing various tasks from product launches to financial analysis, rather than merely answering questions like a typical chatbot.

Chinese officials haven’t specified which regulations the deal broke, but the action seems to serve as a warning to other rising Chinese AI companies against taking their technology outside the country.

According to Han Sheng Lin, a director at Asian Group, “Beijing has drawn a clear line indicating that Chinese AI talent and technology aren’t available for American firms.”

This move effectively put an end to the practice referred to as “Singapore washing,” where several Chinese tech firms, like Shein and TikTok, shifted their headquarters to Singapore to appear less tied to China.

Under Chinese law, all companies must cooperate with the government’s intelligence and security services. Thus, relocating to Singapore was an attempt to convince customers and investors that they were no longer subject to Communist Party obligations—a notion that was always somewhat naive.

A recent hearing by the U.S.-China Economic Security Review Commission suggested that China is actively gathering U.S. data to develop “AI-enhanced intelligence and targeting capabilities for economic competition, political leverage, and wartime advantages.”

“Pigeon” Don

While there is strong bipartisan support in Congress for stricter export controls, President Trump appears to adopt a more pacifist stance.

At this week’s long-awaited summit—postponed from March due to the Iran conflict—Trump may think his options are limited, showing little desire for renewed trade clashes that could further upset the global market.

Last year, Trump rolled back tariffs and lifted restrictions on certain Chinese companies after Beijing threatened to limit access to critical rare earth elements, on which many global tech sectors heavily depend.

Despite the growing evidence of AI-related misconduct, Trump is delivering mixed signals. He has eased regulations on certain NVIDIA chips and indicated that Chinese automakers could feasibly manufacture vehicles in the U.S. amid safety concerns surrounding Chinese electric cars.

Furthermore, his traditionally harsh rhetoric about the long-standing partnership starkly contrasts with his more amicable approach toward Xi. Trump even downplayed reports claiming Beijing might be supplying Iran with resources, affirming on Truth Social, “China has agreed not to send weapons to Iran… We are working together very well! Isn’t that better than fighting?”

By the end of the previous year, approximately one-third of AI models downloaded globally originated from China. Xi is also benefiting from a trend labeled “China Max Shin,” which ignites a frenzy for Chinese tech largely driven by Western influencers, who sometimes seem under the sway of the CCP.

Claims about China’s innovation capabilities and growing acceptance of AI are often echoed by credible Western analysts. Yet, the real story is more complicated, particularly with the CCP’s continuing reliance on grand theft.

Take DeepSeek, a Chinese firm that stunned many by creating an AI model nearly on par with leading Western models but at a considerably lower cost. Last week, its latest iteration drew a collective shrug; production was actually more costly, and the process reportedly involved much more intervention from the CCP.

Delays in utilizing Chinese chips due to CCP constraints meant they eventually reverted to using NVIDIA chips when Huawei models fell short.

mysterious robot

Humanoid robots, often showcased as a breakthrough technology, have performed kung fu at New Year festivities and even participated in the Beijing Marathon. While entertaining, experts remain skeptical about their practical applications.

These ventures symbolically represent the inefficiencies inherent in state-led programs and offer a glimpse into broader themes of Chinese innovation.

For Xi, security and control are paramount. He actively sends out operatives to acquire knowledge and chips, yet any algorithms produced in China must align with the Party’s agenda. There are increasing concerns within the CCP regarding safety, cybersecurity, and employment implications as a result of these technological advancements.

Take, for instance, the recent incident in Wuhan where 200 robotaxis malfunctioned, blocking the streets—a clear wake-up call. At the same time, a report from a government think tank indicates that Chinese workers are becoming increasingly apprehensive about job security in the face of advancing technology.

“We must take decisive actions and carefully anticipate and avert potential problems,” stated Xi.

With youth unemployment hitting around 17%, the situation is particularly dire for younger generations.

The CCP is especially troubled by the phenomenon known as “tanping” or “laying down,” where young individuals withdraw from the harsh labor market to pursue simpler lives.

Recently, China’s Ministry of State Security labeled Tang Ping as a foreign plot aimed at undermining the youth and society. Even as the AI race accelerates, the CCP’s foundational instincts remain eerily unchanged.

Facebook
Twitter
LinkedIn
Reddit
Telegram
WhatsApp

Related News