Opinion | Amar Diwakar

The coming AI Cold War

A screen shows an artificial intelligence news anchor introducing himself at the Light of Internet Expo in November in Wuzhen, China.
A screen shows an artificial intelligence news anchor introducing himself at the Light of Internet Expo in November in Wuzhen, China.AFP/Getty Images

Whatever you think of Trump administration’s blundering trade war with China, it has an objective purpose: to preserve US profits and jobs in high-tech production and services. By most key measures, the US retains its competitive advantage in knowledge-intensive fields, accounting for around 30 percent of global R&D investment and over 45 percent of patents granted.

The Chinese though, have gained ground in a short period. While those advances have been in the medium value-added goods sectors, the United States’ share of world exports in high-tech goods has shrunk as China’s has grown.

Given Beijing’s rapid progression up the tech value-chain, Trump’s gripe that Chinese courts and regulators collude with companies to steal US technology is a legitimate one. However, Trump’s strategy of clamping down on China’s intellectual property theft misses the scope of a crucial future challenge facing US hegemony: the acceleration of Chinese artificial intelligence.

Taiwanese venture capitalist Kai-Fu Lee puts Silicon Valley on notice for failing to account for the gritty gladiators of Sino-innovation in his book “AI Superpowers.” According to Lee, the Chinese are positioned to dominate the AI era thanks to richer datasets and a robust AI-friendly infrastructure. In 2017, Chinese AI start-ups attracted more venture capital than their US rivals did, and PricewaterhouseCoopers estimates that, by 2030, China will add $7 trillion to its GDP as a result of AI deployment. If these trends continue, China’s techno-authoritarianism is poised to win the AI arms race — the stakes of which are enormous.


China’s “Sputnik moment” came in May 2017 when AlphaGo, an AI system developed by British firm DeepMind, defeated Go world champion Ke Jie. Within two months, Beijing unveiled its “Next Generation Artificial Intelligence Development Plan,” which laid out a strategy to build a $150 billion AI industry by 2030. By October, President Xi had designated AI, big data, and the Internet as the trifecta that would transform China into an advanced economy. In a decisive span of a few months, the government had set in motion a new vision of the future. For Lee, “the government’s AI plan was like President John F. Kennedy’s landmark speech calling for America to land a man on the moon.” Meanwhile, Washington, if slow to reckon with the significance of Beijing’s pivot, has been quick to intensify mercantilist bluster in place of a new Chimerican grand strategy.


The paradox of China’s authoritarianism is that it is both repressive and responsive; rather than stymieing Chinese innovation, it has nurtured it. As the dawn of a new era of the digital revolution emerges, the ideological trappings of a new Cold War could pit rival geopolitical blocs against each other while isolating Chinese and American tech sectors from each other, which would paradoxically starve US innovation of the Chinese market it depends heavily on for profits and technical talent. Trump’s bellicosity toward China’s tech ambitions provides an impetus for Xi to double down on his import-substitution drive (“Made in China 2025”) to lessen Chinese reliance on Western technology.

Power in today’s digital ecosystem stems from the accumulation data and weaponizing it into mechanisms of control. As we move into the AI era, data accumulation will be even more extensive along with the leverage accrued to aggregators. This is where China has a decisive advantage in deepening the economic impact of AI: the enormous customer bases for Chinese services mean that the tech sector has more data to use for machine learning, and therefore its algorithms become “smarter” faster.


Eventually, Lee envisions a duopolistic security bloc forming. The existing economic model of developing countries based on low-cost labor will collapse, as the world gradually devolves into a neo-imperial order. Keeping up will require vital investment in digital infrastructure and applications, and thus, a commitment to one of the AI superpowers: either you use Google, buy Apple phones, and drive Teslas; or their counterparts built by Alibaba and Tencent, connect to Huawei and ZTE’s 5G network, and drive Baidu’s autonomous cars.

As Washington and Beijing’s 40-year bargain enters its decoupling stage, a looming techno-geopolitics becomes an ideologically fraught horizon. For if Chinese AI outdoes Silicon Valley, then the future is likely to be entwined with China’s fiber-optic cables and networks, making the export of digital totalitarianism more appealing over its democratic alternative.

Amar Diwakar is a freelance writer.