Will China set up a new state-owned company to monopolize artificial intelligence? – The diplomat


With the recent release of models for major languages ​​like ChatGPT, artificial intelligence (AI) capability has taken a leap forward and garnered significant attention around the world. Inspired by the success of ChatGPT, many Chinese tech companies like Baidu rushed to announce their own plans to develop a Chinese version of ChatGPT. However, to everyone’s surprise, the Chinese government recently banned tech companies from offering ChatGPT-like services and may impose further regulations on AI development.

As AI has gradually become a fundamental part of societal infrastructure essential to national interests, China may set up a new state-owned enterprise (SOE) to monopolize the AI ​​foundation in China, similar to how SOEs monopolize the energy and energy sectors Monopolize the telecom sector.

Traditionally, China’s state-owned companies have controlled industries deemed vital to the national interest and the Chinese economy. It is estimated that SOEs account for at least 23 percent of China’s GDP. Especially as China transitions from an investment-driven export economy to an innovation-driven economy dependent on domestic consumption, the role of SOEs is becoming increasingly important as these SOEs are driving China’s economic transformation. Furthermore, a study indicates that Chinese SOEs form an integrated system that not only benefits the state economically, but also guarantees that the whole country remains under state control.

Coming back to AI, the field is estimated to completely revolutionize the global economy with an estimated contribution of $15.7 trillion to the global economy by 2030. Along the way, AI will become ubiquitous in people’s daily lives. Therefore, AI will not only have a significant economic impact, but will also become essential for controlling the state, fulfilling both key factors for the creation of an SOE.

READ :  How AI is changing intelligent inventory management

Diplomat BriefWeekly NewsletterN

Catch up on the story of the week and develop stories to watch across Asia Pacific.

Get the newsletter

AI has gradually become an essential part of society’s infrastructure in recent years. For example, the Estonian government has launched a new AI-based virtual assistant to give Estonians a voice-based way to navigate through key government services like renewing a passport or applying for social benefits. In Finland, a similar AI platform can offer medical services to help citizens renew prescriptions and inform them about potential health risks. With AI being able to pull, aggregate and analyze sensitive personal data on the entire Chinese population, I find it hard to believe that the Chinese government will leave this capability to private sector companies.

Do you like this article? Click here to subscribe for full access. Only $5 per month.

Aside from AI being essential to the national interest, large AI models like ChatGPT are also extremely expensive to develop and maintain. It has been estimated that each large language model training could cost millions of dollars. On the way to becoming the leading provider of AI technologies in China, venture capitalists and tech companies could easily pour tens of billions of dollars into developing large language models that are very similar to each other, leading to an oversupply of similar AI technologies.

According to the Chinese government, the oversupply that has occurred in the intelligent electric vehicle industry is not an efficient use of China’s financial resources and should not be encouraged. Instead, China can allocate financial and computer resources centrally to one or a few selected institutions to develop basic AI models.

READ :  'The Opportunity is Now': Banks Hiring Amid Big-Tech Layoffs

In the future, the AI ​​ecosystem will be divided into the development of basic models and the fine-tuning of special fields. For reasons of national security and cost-efficiency, I assume that China’s basic models will only be trained by one or a few SOEs with significant financial and computational resources, and most applications will be achieved by slightly fine-tuning these basic models towards specialized applications. such as education, health advice, legal aid, customer service and much more.


Therefore, tech companies in China should focus on leveraging their vertical expertise and know-how, as well as user data, to optimize incorporation models for commercialization in different business verticals, rather than investing an enormous amount of resources in incubation model development.

The creation of a new AI-focused SOE can promote the rapid growth and development of China’s semiconductor industry, especially China’s graphics processing units (GPUs). GPUs are essential for training large language models, but the United States has banned the export of Nvidia’s and AMD’s flagship artificial intelligence chips to China. If AI base models were to be developed by one or more Chinese SOEs, I foresee further tightening of AI processor export controls, and SOEs would have no choice but to rely solely on domestic semiconductor suppliers. This could bring short-term pain but long-term benefits to the entire Chinese AI ecosystem.

To illustrate the intelligence that models like ChatGPT provide, I asked ChatGPT the same question discussed in this article. Below is the insightful and in-depth analysis it has provided. Reading this answer makes me even more convinced that China will create a new SOE to provide fundamental AI capabilities that drive the development of various sectors in China.

READ :  Subaru's EyeSight to incorporate AI; goal is safety, not autonomous driving