Exclusive | The Tech Illusion, DeepSeek, and Taiwan's Semiconductors — Why New Technology Cannot Liberate China, and May Make Taiwan a Bigger Target

From the fax machine to the internet to artificial intelligence, a recurring hope has run through Western democratic thinking: that technology would eventually make China freer. It never has.  Yasheng Huang (黃亞生), a globally recognized authority on China and professor at the MIT Sloan School of Management, told The Storm Media in an exclusive interview that technology has never been a natural ally of freedom — the Chinese Communist Party has not merely adapted to new technologies, but has used them to reinforce control. Taiwan, he warned, now faces a critically underestimated double bind in this technological contest: its semiconductor dominance functions simultaneously as a shield and a target.

The Technology Illusion: From Gunpowder to AI, Authoritarian Systems Have Always Been Early Adopters

"Generation after generation of Western intellectuals — liberal and conservative alike, with no real difference between them — have believed that new technology would produce a freer China," Huang said. "But the Chinese Communist Party has repeatedly shown sufficient flexibility to adapt to, and even co-opt, these technologies in service of its own political agenda and ideological control." （Related： Exclusive | Personalistic Rule, the Hukou Wall — and Why Taiwan's Strongest Weapon Is Itself ｜ Latest ）

This is not an anomaly. In his book《 The Rise and Fall of the EAST: How Exams, Autocracy, Stability, and Technology Brought China Success, and Why They Might Lead to Its Decline》, Huang argues that China invented gunpowder, the compass, and the printing press — and none of them moved the country toward democracy. The political effect of any technology, he contends, has always depended on who controls it and within what institutional environment it is deployed. In China, the internet produced the world's most sophisticated censorship apparatus; social media became an instrument of real-time surveillance and narrative management. This is a systemic pattern, not a policy failure.

AI as Digital Civil Examination: Enforcing Conformity, Not Unlocking Creativity

Artificial intelligence, in Huang's analysis, follows the same pattern. He draws a pointed historical analogy: the way the CCP deploys AI today is functionally equivalent to the imperial examination system. The imperial exams were not designed to identify the most creative minds, but to select those most capable of internalizing and reproducing official doctrine.

"The examination system was an investment in human capital, ensuring you would automatically defer to the autocrat's views; surveillance ensures you won't deviate, because you know your every move is being monitored," he said. Today's China operates both simultaneously — from elementary-school patriotic education to comprehensive data surveillance — forming a precisely engineered dual lock-in mechanism.

The cost, Huang argues, is a systemic hollowing-out of innovative capacity. "China has now reached a point where you cannot even say 'the Chinese economy is not performing well,'" he said. "So who exactly is going to make an accurate diagnosis of the economy?" （Related： Exclusive | Personalistic Rule, the Hukou Wall — and Why Taiwan's Strongest Weapon Is Itself ｜ Latest ）

DeepSeek Rattled the West — But It Proves the Opposite of What Many Claimed

The consequences of that ideological constraint are also visible in the realities of AI competition. When DeepSeek demonstrated near-frontier AI capabilities on mid-range hardware, the immediate reaction in many quarters was that American chip export controls had failed.

Huang disputes this. He identifies two distinct — and mutually contradictory — critiques of export controls that are frequently conflated: the first holds that controls cannot be enforced and China obtains top-tier chips regardless; the second holds that even if controls were strictly enforced, they would not stop China from advancing in AI.

"People sometimes make both arguments at once, without realizing they contradict each other," Huang said. "The DeepSeek case confirms the first critique, not the second. DeepSeek used the most advanced chips Nvidia produces — purchased in large quantities before export controls took effect. It also relied on open-source models, and there are allegations it used OpenAI as a data source. How can you say DeepSeek refutes export controls? Its success was built on international collaboration." （Related： Exclusive | Personalistic Rule, the Hukou Wall — and Why Taiwan's Strongest Weapon Is Itself ｜ Latest ）

What DeepSeek demonstrates, in Huang's reading, is not that controls are ineffective but that they were never rigorously enforced. That distinction has direct consequences for how policy conclusions should be drawn.

At the Negotiating Table, Beijing Cares More About Chips Than Tariffs

When The Storm Media asked whether the West's current export control and friend-shoring strategies need to be rethought, Huang reframed the question at its foundation: if DeepSeek's success confirms that controls were never strictly implemented, the correct conclusion is to strengthen them, not abandon them. "If I were designing export control policy, I would say: we need tighter controls," he said.

Beijing's own behavior at the negotiating table reinforces that assessment. "China's leadership, in negotiations with the U.S. government, has focused far more attention on export controls than on tariffs," Huang said. "That's actually quite revealing, because tariffs affect China's exports broadly — not just technology. Yet they chose to make chips the priority."

That prioritization, he argues, reflects a candid internal judgment. "It signals that they don't believe they can do this on their own," Huang said.

Subsequent developments have sharpened the picture further. Huang noted that Nvidia chief executive Jensen Huang had announced a major order for H200 chips from China, and that Beijing's negotiating position had been structured accordingly: offers to purchase soybeans and aircraft, with the substantive condition being that Nvidia be permitted to sell its most advanced chips to China. "I think Beijing prefers Trump as U.S. president because they believe he can be bought," Huang said. （Related： Exclusive | Personalistic Rule, the Hukou Wall — and Why Taiwan's Strongest Weapon Is Itself ｜ Latest ）

The implications for Taiwan are concrete: American technology control policy — both at the level of enforcement and political will — contains openings that Beijing can probe and exploit. Taiwan cannot treat "U.S. controls plus friend-shoring" as a comprehensive solution. Controls are a tool, not a strategy.

Taiwan's Semiconductors: Shield or Target?

The gaps in export control enforcement leave Beijing in persistent pursuit of the most advanced chip technology — and that pursuit ultimately points at Taiwan.

TSMC and Taiwan's broader semiconductor industry have long been described as the "silicon shield." The phrase captures something real: Taiwan's irreplaceability in advanced-node manufacturing gives the United States, Japan, and Europe powerful incentives to preserve stability in the Taiwan Strait. That is a genuine form of protection. （Related： Exclusive | Personalistic Rule, the Hukou Wall — and Why Taiwan's Strongest Weapon Is Itself ｜ Latest ）

But protection has a corresponding risk. Whether through military coercion, gray-zone operations, supply-chain penetration, or forced technology transfer, the existence of TSMC gives Taiwan enormous strategic leverage in peacetime — while making it the highest-priority target for attack or paralysis in a conflict scenario. Shield and bullseye have always been two sides of the same coin.

A deeper blind spot lies in structural positioning. Taiwan has long cast itself as a technology supplier, and is nearly absent from international debates over AI governance, data sovereignty, and algorithmic regulation. Taiwan sits at the center of the chip manufacturing value chain, yet holds disproportionately little voice in the political discussions that determine how those chips are used, by whom, and under what rules. The result is that Taiwan bears the greatest geopolitical risk while having secured correspondingly little rule-setting influence.

Equally, insufficient understanding of China's military-civil fusion mechanisms constitutes a structural vulnerability for Taiwan. Civilian AI, drones, and satellite communications are being simultaneously converted into military capabilities under the military-civil fusion framework. If Taiwan continues to view cross-strait technology interactions through the lens of technological neutrality, it is applying peacetime logic against an adversary that has already shifted to a wartime mindset.

No Single Industry Can Sustain a Nation

When Huang addressed the structural problems in China's industrial policy, he invoked Taiwan explicitly as a point of comparison. "Even for a relatively smaller economy — like Taiwan — you cannot rely on a single industry," he said. "You need a diversified economic structure, different industrial sectors. The problem with the scale-up model driven by industrial policy is that it tends to focus only on technology. And the long-term costs of that are extremely high."

The statement was directed at China, but the warning applies equally to Taiwan. If the "silicon shield" narrative hardens into structural over-dependence on a single sector, Taiwan risks replicating — on a different register — the very mistake China is making: using scale to mask a shrinking range, and using the peak of one industry to obscure the erosion of broader economic diversity.

Technology did not automatically liberalize China. It will not automatically protect Taiwan. From gunpowder to artificial intelligence, the political direction of any technology has always been determined by the institutions that deploy it and the strategic will behind those institutions — not by the technology itself.

（Related： Exclusive | Personalistic Rule, the Hukou Wall — and Why Taiwan's Strongest Weapon Is Itself ｜ Latest ）

​