AI has moved squarely into the forefront of everyone’s lives, whether we realize it or not. The technology promises “epoch-making transformations—in society, economics, politics, and foreign policy” (Kissinger 3), in addition to already powering research and services used by billions every day. In this new landscape, “data is rocket fuel for AI,” thus creating a strong feedback loop between collecting user data and gaining more users (Smith 284). This situation means that first-movers in the field—namely the US and China—have a massive advantage when it comes to AI, leaving other countries to “pick up the scraps” (Lee 169).
However, although the United States benefitted from an early start in AI, China has rapidly made up lost ground, and is now “the only true national counterweight to the United States in this emerging technology,” with no signs of slowing down (Lee x). If we continue our current trajectory with AI research, China’s competitive advantage in AI has massive negative implications for the future of American relative political and economic dominance, in addition to posing a grave threat to American constitutional democracy.
It is important to briefly highlight AI-adjacent areas where China shines, and how this gives the country a massive comparative advantage over the United States. First and foremost, the Chinese government has created a much friendlier policy environment for AI than their American counterparts across the Pacific. Where American politicians pump the breaks on government support for AI, the Chinese push down the accelerator. Central government promises of funding and support for AI research have “acted as a beacon to local governments throughout the country to follow suit” (Lee 17). Compare this to the United States, whose “combative political system aggressively punishes missteps or waste in funding technological upgrades,” (85), thus creating resistance to government-spearheaded AI investment and research. Lastly, the ever-taller Great Firewall of China effectively excludes foreign competition from operating in the country due to laws requiring all firms to “provide access, cooperation, or support for Beijing’s intelligence-gathering activities” (Tanner).
While closing off their country to foreign firms may appear to be stifling competition in exchange for control, the norms and values driving Chinese tech entrepreneurship ensure a ruthlessly competitive and innovative environment for AI development. Kai-Fu Lee characterizes the Chinese tech market as “a coliseum where hundreds of copycat gladiators [fight] to the death” (Lee 24). In the Chinese market, the best ideas—Western or Chinese—are immediately copied and iterated, to hold onto precarious market share. In Silicon Valley, copying is seen as stifling innovation and the very thing that would hold China back. However, this is not the case for two reasons:
- Adaptive Optimization: Pure copying does not ensure success in the ruthless Chinese market—it is merely the first phase. After taking an American business model, Chinese firms run with it and specifically optimize it for Chinese users (24).
- Capital Redistribution: Chinese firms save capital by copying algorithms and research from the West and pour this money into applications (NSCAI 161).
Coupled with intense local competition, this means that legwork on AI research done by the West is used to elevate the level of innovation in the closed-off Chinese market—innovation that has led Chinese companies to hold three spots with ‘The Seven Giants’ of AI research (Lee 91).
Lastly, with both the largest population in the world and a surveillance-oriented authoritarian government, China has a truly staggering amount of data available at its fingertips. Thanks to the friendly policy environment, Chinese companies have unparalleled access to user data, access which “vastly outstrips what Silicon Valley’s leading companies can decipher from [one’s] searches, ‘likes,’ or occasional online purchases” (Lee 17). As data creates user feedback loops that strengthens AI’s effectiveness, China has the potential to stand head and shoulders above the rest of the world in AI in short order. The final advantage to China in this regard is data’s outsized importance to AI, as—at least for now—data-driven AI implementation is more important than open-ended research and discovery.
The consequences of China’s AI advantage have the potential to be severe for the United States:
- Economic Shift: From an economic standpoint, China’s “lead in AI development will translate into productivity gains on a scale not seen since the Industrial Revolution” (Lee 18). With stronger algorithms on their side, China may corner the market for commercial AI applications, reaping immense profits in the process. At the same time, while the US may dominate AI design, China can better use current AI processes to dominate discoveries in nearly every other field.
- Brain Drain: Critically, China’s national push for AI development and AI-friendly policy also has the potential to create a brain-drain and influx of AI-experts into the country, as the wealth and data of top AI companies attracts top AI talent (Lee 20). This means that the US’ supply of AI experts may soon run out, and that even America’s current edge in AI-design is not safe.
- Democratic Vulnerability: Furthermore, China’s AI advantage poses a significant threat to American constitutional democracy. The National Security Commission on Artificial Intelligence (NSCAI) states that “[w]herever China controls the digital infrastructure, social media platforms, and e-commerce, it would possess greater leverage and power to coerce, propagandize, and shape the world to conform to its goals” (26).
This structural leverage is already playing out in the social media sphere: TikTok, one of the most ubiquitous apps in the US today, is owned by ByteDance—a Chinese company, and the app’s privacy policy states that it collects user data to customize content, inform its algorithms, and infer information about the user, such as age, gender, interests, and more (TikTok Inc). Recalling the close relationship between Chinese tech companies and the Chinese government, this means that China has access to tens of millions of Americans’ data and may even be able to exert influence on the algorithms which determine what content TikTok users see.
Data “may be our era’s microscope or telescope—making possible important, even revolutionary insights” (Stephens-Davidowitz 16). As it stands, China’s AI advantage will allow ever greater aggregation of American data. The long shadow of Russian influence in the 2016 presidential election still looms large in the US, and with political competition with China increasing year by year, China is in a prime position to exploit its AI advantage and infrastructure against American democracy.
However, the picture does not need to be this grim. While holding the definitive comparative advantage in AI may be out of reach, this does not have to come at the cost of American prosperity or constitutional democracy. The NSCAI says AI “will be the most powerful tool in generations for benefiting humanity” (NSCAI 1). There is so much good and potential in AI for helping humanity—whether in developing pharmaceuticals, diagnosing patients, or unlocking the mysteries of the natural world. But it depends on the US to maximize these benefits while mitigating the risk from Chinese AI. Brad Smith suggests we pursue the following course: “we need to democratize AI and the data on which it relies” (Smith 286).
Democratizing is not as far-fetched as it may seem. Firms such as OpenAI are breaking major ground, with their GPT-3 algorithm API open to the public (286). Moreover, it is in the best interests of all AI researchers—including those in China—to rapidly share data and algorithms for AI, for the sake of comparing and improving each other’s work (Lee 87). Furthermore, in the same fashion that China uses AI to pursue authoritarian goals, democracies can “ensure that the use of AI by the government is limited by wise restraints to comport with the rights and liberties that define a free and open society” (NSCAI 27). Democratizing AI would take away the danger of consolidated data and control, encourage an open forum for discussion of AI applications, and allow humanity to continue benefiting from AI’s advancements.
China is on a path to AI dominance, and in our current situation, this has grave consequences for the United States and the free world. We are at a crossroads in the US both technologically (in nearing our full capabilities in AI research) and ideologically (in figuring out how this affects our democracy, and what to do about it). AI has both positive and negative impacts, and it cannot be said whether the technology on the whole is ‘good’ or ‘bad’. What can be said is that this is our decision to make. The current status quo will not serve our interests; we need “White House leadership, Cabinet-member action, and bipartisan Congressional support to win the AI era” (NSCAI 8). AI regulation must enshrine freedom and democratic values and turn the technology towards this goal. There is no one right approach to consider, but American action must be taken; for “[w]hen it’s wrong to push for regulation, only the wrong will get regulation” (Lessig 338).