China's AI Strategy: Profit-Driven Industrial Revolution

At first glance, AI’s use cases in China stand in contrast to the typical Western conceptualization of AI. In China, the “world’s factory,” ML, and automated quality control – not LLMs – are leveraged alongside profit-seeking motives to drive application in the raw materials and manufacturing sectors. These are done with the goal of enhancing both quality and efficiency to directly increase profits. 

In the West, on the other hand, novel LLMs appear to still be having their moment in the sun, where they are marketed as “co-pilots,” AI companions, or personal assistants – automating mundane service sector tasks through one-off prompts. This differs from the seamless and constant integration of algorithms observed in automating industrial sector tasks in China. 

Statements and interviews from government officials, along with proactive initiatives by Chinese entrepreneurs further indicate a growing emphasis on utilizing and enhancing AI for profitability in the raw materials and manufacturing sectors, rather than merely assisting with tasks like email composition or customer service in the service industry. 

As mentioned in the review of the first half of the year, these initiatives are quickly coalescing into a capable whole-of-government approach – focused on hard, industrial AI applications, not soft, theoretical developments – an approach whose adoption rate and speed has not yet been seen in other sectors or geographies. Whether it will result in the dividends needed to jumpstart China’s economy is still yet to be seen.

This shift, from a cutesy chatbot for Chinese netizens to a sleepless algorithm aimed at enhancing efficiency in raw materials and manufacturing, must be scrutinized in light of China's struggling post-COVID economy. Recent reports on the emergence of "new forces of production" and the industries targeted by this development underscore the necessity for China to harness AI effectively and generate sufficient value to revive its economy.

Since the formal termination of China's Zero-COVID Policy in January 2023, the nation has faced persistent challenges, including sluggish economic growth, overcapacity concerns, and internal issues such as an aging population and mounting local government debt. These factors have led some economists to suggest that China may have become ensnared in what is known as the “middle-income trap.”

This trap, caused by an inability to reform market forces to attain the “innovative capability needed to boost productivity and compete with developed countries,” has been overcome by China’s Asian neighbors – like Japan, South Korea, Singapore, and Taiwan – primarily through technological innovation. When done right, innovation allows these developing economies to secure pieces of the “advanced economy” pie, creating higher-value goods and commanding higher prices within the market – thus escaping the middle-income trap.

In our previous analysis of China, we emphasized the challenges it faced due to limited resources such as human capital, semiconductors, and rare earths as potential impediments to domestic AI development. However, recent developments in advanced packaging, rare earths export restrictions, and advanced semiconductor manufacturing (even if yields are low) suggest that China's whole-of-government approach is becoming more adept at employing workarounds to overcome these obstacles and wiggle its way out of the middle income ensnarement. The pivotal question lies in solvency: Will these advancements enhance the value of Chinese goods or simply inundate the market with cheaper alternatives? The outcome will effectively determine whether China propels itself further into the middle-income trap or follows the trajectory of Korea and Japan in breaking free from its constraints.

In August 2023, China made history by pioneering legislation specifically tailored for GenAI, a response triggered by reverberations stemming from the introduction of technologies like ChatGPT. Shortly after placing that piece into AI's legal framework in China, policy documents pertaining to “strategic emerging industries” began to increasingly incorporate AI.

Strategic emerging industries in China include AI; Information technology; Biotechnology; New energy; New materials; High-end equipment; and Green environmental protection. Although AI emerges as a strategic industry in its own right, other designated sectors boast high translatability, facilitating the seamless integration of AI technologies. This adaptability has already sparked significant developments across various sectors, a few of which are outlined below.

Biotechnology

New energy (vehicles)

New materials

Green environmental protection

China's dominance in green energy and AI advancements ensures continued and expanding AI applications across solar, wave, tidal, and other sectors.

Navigating the regulatory landscape is becoming increasingly complex for publicly-facing LLMs in China. This raises concerns about the feasibility of censoring algorithms, which are typically based on Western models subject to less censorship.

However, while challenges abound for publicly-facing LLMs, fewer barriers appear to exist in the commercial and industrial sectors. Chinese policymakers recognize this by integrating AI into various industrial policy documents, emphasizing its application as a “new productive force” within “strategic emerging industries.”

Despite potential geopolitical scrutiny, partnerships like Nvidia's collaboration with Chinese EV manufacturers – apparently unconstrained by export controls – indicate that market dynamics often outweigh policy constraints.

Although challenges such as human capital, semiconductor production, and rare earths will likely remain as significant perennial challenges, China's deep state coffers appear capable of mitigating their impact. While these challenges may impede efficient development, they are unlikely to halt China's overall advancement.

Q3 2023: July

Amidst a sweltering Chinese summer, Shanghai played host to the World Artificial Intelligence Conference, which brought tech giants like Baidu, Huawei, 360 Security Technology, and Alibaba together. There they announced that they would be spearheading a task force to shape a new, Chinese LLM standard.  

About a week later, the Chinese government published its finalized rules on generative artificial intelligence.

September

Amid the ongoing excitement over China's booming AI sector, the Ministry of Industry and Information Technology, alongside other key ministries and commissions, unveiled an ambitious "Three-Year Action Plan for the Industrial Innovation and Development of the Metaverse (2023-2025)." Signaling a concerted effort to integrate emerging AI technologies not only into China's industrial framework but also into service sectors like education, culture, tourism, and media. This expansion beyond raw materials and manufacturing reflects China's recognition of the need to stimulate investment and innovation in services. However, the implementation and potential impact of this plan remain unclear amidst the plethora of national-level documents published annually.

Q4 2023: October

In mid-October, China, recognizing its first-mover advantage when it comes to legislating AI, the Cyberspace Administration of China released a statement proposing Global Artificial Intelligence (AI) Governance Initiative. It recommends enhancing laws and regulations for safeguarding privacy and data security, and preventing misuse. The release also calls for human, risk-based testing of AI systems to be put in place as well.

Likely realizing Chinese capabilities in AI were and are being made possible by US chip inputs, the US Department of Commerce tightened export controls on chips. According to synthesis by Georgetown’s Center for Security and Emerging Technology, “These export control updates are intended to stay up-to-date with technological developments of AI systems and manufacturing of advanced chips.” China is unlikely to say outright if or how these updates might be affecting their AI capabilities, but be on the lookout for a slow in title page news about Chinese capabilities – as the absence may mean the US’s bigger yard, higher fence renovations are working.

November

In November Ant Group, the financial technology affiliate of one of China’s largest companies, Alibaba, received Chinese government approval to release products powered by its "Bailing" AI LLM. As mentioned above, considering China’s relatively mature legislation on all-things-AI, obtaining regulatory clearance is a significant hurdle that operators must clear before offering their products to the public.

Later that month, China's AI sector began signaling a decline in investor interest due to “challenges in technology commercialization, unclear business models, and concerns over investment returns.” In 2023, venture investments shifted towards LLMs, robotics, machine vision, and AI chips.

Notable funding rounds this year included Zhipu AI's $345 million financing and Baichuan Intelligence's $300 million USD Series A1 round. 01.AI, founded by Lee Kai-fu, hit a US$1 billion valuation after a funding round including Alibaba's cloud unit. Robotics and machine vision industries have also attracted significant investments, with notable companies like Zhiyuan Robotics and Suiyuan Technology raising substantial funding rounds. 

December

In December, Shanghai played host to the Embedded AI Summit, yet another China-hosted summit which gathered leading academics, chip manufacturers, and aerospace IT technology providers to speak on questions like smart buildings, AIoT for indoor vertical farming, and the perennial issue of bad data for AI providers.

As China progresses, it moves six months closer to its goal of "becoming a global AI leader by 2030." The transition from the first half of 2023 to the second half is noteworthy. Initially, there was a proliferation of different types of LLMs, accompanied by governmental attempts to regulate their usage. However, decision-makers have since recognized that the true value of AI for China lies not in the service sector, as in the West, but perhaps in its utility for raw materials and manufacturing. Its application potentially even holds the key to jumpstart its post-COVID economy and getting out of the middle-income trap. This shift underscores a strategic reorientation towards leveraging AI's potential in industrial applications to drive economic growth.

While initial movements are being made within metaverse ventures to jumpstart Chinese AI activity within the service sector, it is still unclear whether these will have the effect AI appears to be having within industry.

We’re likely to see much more innovation within Chinese greentech – both in terms of hardware and software – and considering the slew of rapidly approaching climate deadlines, this may just be a piece of the international economy China needs to secure to climb the high-quality and high-value ladder and turn itself into a high-income economy.