ChatGPT, GPT4 and other “big model” applications in artificial intelligence have rapidly gained popularity, marking a new development stage in the technology.
There are different views on AI, and the future remains uncertain. Others are concerned about the risks of rapid AI development — the technology getting out of control, its impact on job markets and the emergence of new social problems and security risks.
Wide social penetration and the potential value of universal use differentiate ChatGPT and GPT4 from previous AI applications. Some early tests show that the new generation of AI applications has a learning and understanding ability greater than that of most humans. Such applications also show an unprecedented evolutionary speed. In less than five years, the GPT training model has had four iterations, with more than a hundredfold efficiency improvement each time. The speed of technological progress is much faster than Moore’s Law, a doubling every two years.
This unprecedented evolutionary speed has enabled ChatGPT to cross the threshold from testing to practical use in a short period of time. It has quickly become available in thousands of households, and may be applied and expanded in more fields. This speed of progress has also strengthened the double-edged sword effect of AI. The technology not only brings huge business prospects and industrial changes but also may deal a blow to social stability in some populations. This double-edged sword effect, therefore, must be addressed when countries develop a new generation of AI.
In the birthplace of ChatGPT, the American tech community has concerns over the future development of this technology. One group of people, exemplified by Elon Musk, have urged countries to prioritize the protection of human control and restrict the research and development of more advanced AI. The appeal has caused controversy in the United States, but it has nonetheless demonstrated mainstream American society’s knowledge of the the technology’s double-edged effect.
China and other developing countries face an even harder choice. On one hand, they can learn from the experience and lessons of developed countries and avoid detours in industrial development and technical governance. On the other hand, they worry that they will be left behind in the tech revolution and become the earliest victims of technological substitution. In addition, developing countries are not immune from the huge social impact and potential societal security threats brought by AI.
China has prepared answers to these questions. Since 2019, the Chinese government has issued a series of documents on the governance of AI, and established its basic position and view on AI’s development and security. These documents contain some forward-looking and innovative ideas that not only help balance the opportunities and risks of AI in China but also serve to inform more developing countries making the same attempt. The Chinese thinking can be summarized in four principles:
• First, putting people at the center. The fundamental task of AI as an information technology is still to serve the development of human society, not to bring a subversive impact on it. China, the most populous country in the world, gives priority to the social value of this technology. It emphasizes “AI for good,” pursuing the safety, reliability and human control of AI development and preconditioning AI research and promotion on addressing ethical problems first. China will fully consider the ability of society — especially ordinary people — to adapt to the new-generation AI instead of pursuing technological development at the cost of people’s well-being.
• Second, aligning with national development strategies. Countries have different focuses in AI development and industrial layout. China has some technical acumen in autonomous driving, machine vision and smart cities, but it’s still at an early stage in the “big model” path represented by ChatGPT. The emergence of ChatGPT will not change China’s vision of AI industrial development because it advocates the use of AI for quality development and people’s livelihoods. Its existing industrial layout also serves this strategic goal. Whether or not new applications are aligned with China’s development concepts will be the key success factor for them domestically.
• Third, prioritizing regulation and governance. China emphasizes the rule of law, and AI must not be an extrajudicial domain. To ensure that the development of AI conforms to the law and that civil rights are protected, China has existing legal provisions on user privacy, data security, fairness, and transparency in algorithms, and it will strictly regulate data collection and technical application by AI enterprises. In recent years, China has cracked down on some deep fake applications that are harmful to society. In the future, China will not allow AI applications that may endanger national security, social stability or people’s rights to develop in disorder. A full-fledged legal system and governance are the prerequisites for the rapid development of AI applications in China.
• Fourth, promoting the common well-being of all mankind. China stresses that scientific and technological progress should ultimately benefit everyone and it opposes technical barriers or divisions. In this connection, China firmly stands on the side of developing countries and calls on all countries to regard the further AI evolution from the perspective of human society as a whole. To avoid vicious competition in this field, which could lead to technological abuses that endanger human society, China supports strengthening international exchanges and building common rules and a global governance system. China has also called on all countries to pay more attention to improving the adaptability of ordinary people to AI in the course of technological development to promote coordinated development between human society and AI applications.
These four principles are based on the experience of cutting-edge technology governance in China in recent years, and informed by the concerns of ordinary people in various countries over the technological revolution. Cutting-edge technologies, including new-generation AI, should not be divorced from the fundamental purpose of serving society, let alone becoming a tool and platform for a few countries and enterprises to expand their hegemony.
China calls for the development and application of emerging AI technologies with a responsible, open and win-win attitude, believing that this concept can solve the double-edged sword problem and is more in line with the fundamental interests of developing countries and ordinary people around the world.
