How China Regulates AI to Maintain Control

China's AI Ambitions: A Cautious Path to Power

by Faruk Imamovic
How China Regulates AI to Maintain Control
© Getty Images

Since 2017, the Chinese Communist Party (CCP) has meticulously charted its course to dominate the realm of generative artificial intelligence (AI). These sophisticated programs utilize massive datasets to recognize patterns and produce seemingly spontaneous knowledge. By 2020, China aimed to achieve significant advancements in AI models, core devices, high-end equipment, and foundational software. However, the launch of OpenAI's ChatGPT in late 2022 disrupted Beijing's plans, highlighting the United States' leadership in the AI race and exposing the great-power competition to global scrutiny.

America's lead in AI is not guaranteed to last forever. China's national tech champions have entered the fray, molding a technology that thrives on free-flowing information to fit within China's tightly controlled information bubble. While censorship may slow China's AI development and limit the commercialization of domestic models, it will not prevent Beijing from leveraging AI for its strategic interests. China's leader, Xi Jinping, views technology as the key to revitalizing the country's economy. Even if China does not surpass the US in the AI race, securing second place still holds significant power and potential risks.

"There's so much we can do with this technology. Beijing's just not encouraging consumer-facing interactions," said Reva Goujon, a director at the consulting firm Rhodium Group. "Real innovation is happening in China. We're not seeing a huge gap between the models Chinese companies have been able to roll out. It's not like all these tech innovators have disappeared. They're just channeling applications to hard science."

The CCP's Vision: AI as a Tool of Control

In its internal documents, the CCP outlines its intention to use AI to shape reality and strengthen its grip on power within China. AI is poised to play a role in political repression, surveillance, and monitoring dissent. Additionally, the CCP plans to drive breakthroughs in industrial engineering, biotechnology, and other critical fields through AI. Despite lagging behind US tech by a few years, China's AI advancements can still have a profound geopolitical impact.

China's State Council's 2017 Next Generation Artificial Intelligence Development Plan envisaged AI as a means to "grasp group cognition and psychological changes in a timely manner," thereby enhancing social governance and maintaining stability. This reflects the CCP's belief that AI, if designed to their specifications, can fortify their power. This month, the Cyberspace Administration of China, the country's AI regulator, launched a chatbot entirely based on Xi Jinping's political and economic philosophy. Predictably, ChatGPT remains unavailable in China and Hong Kong.

For the CCP, AI offers a new means of mass surveillance and information control, especially crucial given the current economic challenges. The Chinese economy has hit a rough patch, with traditional growth models no longer delivering the desired results. Wall Street, Washington, Brussels, and Berlin have acknowledged that China's economic model is outdated, and Beijing has yet to find an effective replacement. Infrastructure and industrial expansion no longer provide the same economic boost. The world is pushing back against China's exports, and domestic consumption drives have been largely unsuccessful. The property market is distorted, growth has plateaued, and deflation looms. In the face of increasing public dissent, with record numbers of demonstrations against government policies, Xi needs an effective way to maintain control.

Xi Jinping
Xi Jinping© Getty Images

A Double-Edged Sword

China's approach to AI is twofold: harnessing the technology's power while maintaining strict control over its outputs. The CCP has constructed a regulatory framework that is both flexible enough to accommodate large language models and stringent enough to control their outputs. AI models for public consumption must be registered and approved by the Cyberspace Administration of China. This registration includes providing information on training datasets and test results, enabling the CCP to maintain control over AI developments.

"The real challenge of LLMs (large language models) is that they are really the synthesis of two things," said Matt Sheehan, a researcher at the Carnegie Endowment for International Peace. "They might be at the forefront of productivity growth, but they're also fundamentally a content-based system, taking content and spitting out content. And that's something the CCP considers frivolous."

In recent years, the CCP has shown a willingness to cut out technology it deems frivolous or harmful to social cohesion. For instance, in 2021, the CCP barred anyone under 18 from playing video games on weekdays and paused new game approvals for eight months. However, AI is more than just entertainment; it's integral to the future of computation. The CCP cannot ignore the impact of OpenAI's viral success, the implications for the US-China tech competition, or the potential for AI to boost economic growth and political power.

The Global Implications of China's AI Strategy

To control AI effectively, China must develop models that suit its specific purposes. While Chinese tech giants are playing catch-up, they are also making significant strides. Baidu's chatbot, Ernie Bot, boasts 200 million users and 85,000 enterprise clients, and ByteDance's Doubao bot has quickly become the most downloaded in the country. However, the generative AI industry in China is still in its early stages, with companies investing heavily in infrastructure and model training.

The challenge lies in moving from training and tweaking models to making profitable applications. A price war among Chinese companies may drive a race to the bottom, similar to what was seen in the electric vehicle industry. The government initially subsidized design and production, but eventually supported only end consumers, leading to a consolidation of the market. A similar approach in AI could result in a few dominant players, supported by the government but burdened by the CCP's control.

China's AI regulations, such as those for deepfakes and authenticity labeling, highlight the need for accountability in AI development. While the CCP's checks may go too far, the US regulatory framework currently lacks systemization. The Commerce Department's recent initiative to test AI models for safety is a good start, but more comprehensive measures are needed.