TL;DR:
- China seeks to balance innovation and authoritarian control in its pursuit of a hyper-advanced economy.
- The government promotes favored high-tech sectors while tightening control over disapproved industries.
- The tension is most pronounced in the field of artificial intelligence (AI).
- Strict regulations govern consumer-facing generative AI and AI-generated content.
- Enterprise AI faces fewer constraints and is prioritized for strategic development.
- Government-backed investment funds and data exchanges support the growth of AI businesses.
- Challenges include data privacy concerns in industries like self-driving cars.
- China’s approach will shape the future of AI development in the country.
Main AI News:
As China strives to become a hyper-advanced economy, it grapples with the age-old dilemma faced by authoritarian states: how to balance innovation’s growth potential with the need for tight control. President Xi Jinping’s administration actively promotes the commercialization of preferred high-tech sectors, including electric vehicles and quantum computing, while simultaneously clamping down on disfavored industries. The online tutoring sector faced strict regulations in 2021 due to concerns about high tuition fees, and recent measures have limited in-game purchases in the video gaming industry, leading to a significant drop in Tencent’s market value.
This tension is particularly evident in the field of artificial intelligence (AI), a technology seen as vital both economically and strategically worldwide. In Beijing, the concern extends beyond the technology’s potential misuse; it also revolves around AI’s dependence on vast data and unregulated spaces, making it potentially subversive if left unchecked. Hence, China is diligently reinforcing its “great firewall” to safeguard against these risks in the AI age.
Two Decades of Transformation
Bill Clinton, former U.S. President, once likened China’s struggle to control the internet to “trying to nail Jell-O to the wall.” Today, however, the jelly seems firmly in place, as Western internet services like Facebook, Google, and Netflix remain largely inaccessible to the Chinese population. Local platforms rigorously censor undesirable content through algorithms and government oversight. In 2020, a tech crackdown brought powerful Chinese tech giants, Tencent and Alibaba, closer to the government’s influence, with the state taking stakes in these firms and showing keen interest in their daily operations.
This strategy has resulted in a thriving but sanitized digital economy. For example, Tencent’s super-app, WeChat, combines messaging, social media, e-commerce, and payments, generating hundreds of billions of dollars in transactions annually. President Xi hopes to replicate this delicate balance with AI, but skepticism persists among foreign experts.
Building Blocks of Control
China’s approach to AI control begins with the strictest rules worldwide governing Chinese equivalents of AI models like Chatgpt. Since March, companies must register any algorithms that make recommendations or influence people’s decisions with officials. Furthermore, the government issued rules in July mandating that all AI-generated content should uphold socialist values, effectively banning content that challenges the party or ridicules Xi Jinping. In September, a list of 110 registered AI services was published, with the registration process’s details remaining confidential.
In October, a national information security standards committee issued safety guidelines for generative AI models, requiring extensive self-assessment of training data, manual testing of subsets, and the rejection of unacceptable content. Unregistered algorithms are to be blocked, and their creators face penalties. This stringent approach has slowed the adoption of consumer-facing generative AI in China, with some models, like Baidu’s Ernie Bot, experiencing delays in release due to compliance concerns.
Enterprise AI Takes Center Stage
In contrast to consumer AI, enterprise AI faces fewer regulatory constraints in China. This has driven capital and labor towards machine learning for business applications, aligning with the government’s goal of rivaling the United States in AI without the potential subversion of consumer AI.
Cities like Shenzhen have launched substantial AI-focused investment funds to support this strategy. For instance, a 100-billion-yuan ($14 billion) fund was established in Shenzhen, and similar initiatives have emerged in other cities. These funds support companies like Qi An Xin, offering generative AI solutions for data-security management.
Unlocking Corporate Data
Another crucial aspect of China’s strategy is transforming corporate data into a public resource. While the state does not seek ownership, it aims to control data flows. To achieve this, the government promotes data exchanges, enabling businesses to trade standardized information across various commercial sectors. This democratizes access to valuable knowledge previously held by tech giants and provides real-time insights into the economy for banks and brokers.
Approximately 50 data exchanges operate across China, with the Shanghai Data Exchange leading the way. These exchanges offer a wide range of data products, from energy sector information to real-time healthcare data. As these exchanges gain momentum, the datafication of industries could provide a substantial economic boost.
Challenges on the Horizon
However, challenges persist in China’s pursuit of enterprise AI dominance. For instance, the car industry’s transition to self-driving vehicles relies heavily on data. Companies developing self-driving algorithms need extensive data sets for training, but concerns about data privacy and security hinder data sharing. Chinese authorities are cautious about sensitive information falling into the wrong hands, as exemplified by the crackdown on Didi Global in 2021.
To address these concerns, car manufacturers are collaborating with state-backed companies to manage driving data and ensure compliance with local regulations. Some companies have also limited certain features to prevent data misuse. Similar trade-offs between innovation and security are likely to affect other industries, slowing the development of enterprise AI.
Conclusion:
China’s evolving approach to AI control reflects its complex balancing act between promoting innovation and maintaining strict control. While hurdles remain, the country’s determination to assert itself as a leader in enterprise AI and data utilization is evident. How China navigates these challenges will shape the future of AI development in the country.
