Artificial Intelligence (AI) is a buzzword in broader society and also an area of focus in the community of philosophy. Regarding the philosophical study of AI, academia pays great attention to the subjectivity and ethics of AI, as well as its impact on the division of labor and contributions to the emancipation of humanity.

In recent years, research on the philosophy of AI has reached a fever pitch, but this is not the first time China has studied this topic. In fact, China has decades of tradition in AI-related philosophical studies and has experienced shifting research focus throughout different historical periods.

Although the philosophy of AI is closely related to frontiers in science and technology, it is necessary to “look back” to map out its development course while “looking forward.” This is not only essential as we write the contemporary Chinese history of philosophy and thought culture, but can also illuminate understandings of how philosophy can better serve us in the science and technology era.

Origins

How did the philosophy of AI originate in China? There are two clues to the origins of AI itself. First, is the conceptual origins from studying intelligence to artificial intelligence, while the other clue is in the theoretical origins from cybernetics to artificial intelligence.

The emergence of AI, as a concept in China, can be traced back to the 1960s and 1970s. The 1965 book An Introduction to the Disciplines of Natural Sciences is arguably the first relevant source with a third-tier entry labeled “Artificial ‘Intelligence’ Research” under the second-tier entry “Information Processing Theory,” and further under the first-tier entry “Computer Technology.” The Bibliographical Index to Bionics II, published in 1972, made it clear that the Chinese characters “人工智能” (rengong zhineng) is equivalent to “Artificial Intelligence” in English. It is the earliest Chinese document using the exact term “Artificial Intelligence.”

Computer technology was far less developed half a century ago than it is today, but the important development goal—for machines to simulate human intelligence, or the brain—had already been set back then. While iterative computer technology was advancing AI in practice, cybernetics theoretically linked the ideas of “artificial” and “intelligence.”

AI was initially regarded as one of the four branches of cybernetics in their giant science system, with the other three branches being engineering cybernetics, bio-cybernetics, and socio-cybernetics.

Just as there is a linkage effect between scientific research on the “three theories” (system theory, cybernetics, and information theory) and philosophical studies, AI research’s starting point in China was synchronic with related philosophical research.

In the 1960s and 1970s, philosophical research of AI in China was characterized by limited translated works and fragmentary discussions. The field didn’t gain steam until after the reform and opening up.

In the mid to late 1970s, the Journal of the Dialectics of Nature and Selected Translations of Foreign Natural Sciences and Philosophy included translations of research papers authored by scholars from many countries, including the Soviet Union, the United States, Japan, Britain, Germany, and France. However, due to a peculiarity of the times, AI research was teeming with philosophical criticism, such as the opinion “AI has the decaying smell of idealism and a metaphysic world outlook.”

By the end of the 1970s, aside from journals on the dialectics of nature, comprehensive philosophical periodicals such as Philosophical Research and Social Sciences Abroad also incorporated the philosophy of AI into their vision.

On disciplinary levels, early philosophy of AI was inseparable from the development of disciplines concerning the dialectics of nature. For example, the Draft of 12-Year Research Planning (1956–1967) for the Dialectics of Nature on Philosophical Issues in Mathematics and Natural Sciences included “philosophical issues in cybernetics.” The sixth article “Research of Philosophical Issues in Natural Sciences” in the 1978–1985 Draft Planning Outline on the Development of the Dialectics of Nature underscored the need to research philosophical issues in AI, cybernetics, informatics, and bionics.

Research trend in the 1980s

In the 1980s, China saw the first wave of philosophical research on AI. It was promoted by two branches of philosophy: the dialectics of nature, which played the guiding role, and Marxist philosophy, which expanded the research scope for AI philosophy.

The first branch was represented by pioneering scholars like Chen Bu, Tong Tianxiang, Fan Dainian, and Ding You, all from the Department of the Dialectics of Nature at the Institute of Philosophy under the Chinese Academy of Social Sciences (CASS). They published papers, founded journals, held conferences, and established organizations.

In 1981, the Institute of Philosophy at CASS joined forces with the Institute of Automation and the Institute of Psychology at the Chinese Academy of Sciences to hold a seminar on issues related to the philosophy of AI. This represented the field’s first conference in China. Attendees were from different institutions, and covered philosophy, computer science, psychology, linguistics, and informatics. The seminar was heralded as “an alliance of natural scientists and philosophers.”

The same year, the Chinese Association for Artificial Intelligence was inaugurated. Although the association mostly consisted of natural scientists, it was at first affiliated with CASS, with Tong Tianxiang elected as vice chairman, which reflected philosophy’s strong presence in the AI realm.

Meanwhile, scholars of Marxist philosophy paid close attention to new phenomena in AI development. Many textbooks and treatises of Marxist philosophy in the 1980s made special references to the philosophy of AI, marking it as an emerging theme in philosophy, with universality, epochal characteristics, and comprehensiveness.

Featuring numerous achievements of different forms, the wave of research had a focal point: epistemology, which can be subdivided into three themes.

The first, and primary theme, was the relationship between mind and matter. Machines’ simulation of the human brain, or the mind, was the original issue in AI philosophy. In the 1980s, theoretical and practical explorations of AI went hand in hand, mainly concerning natural language processing, theorem proving, manufacturing automation, and expert advisory systems. Based on the characteristics of human thinking, AI was supposed to face two difficulties: recognizing the brain as an organ for thinking, and equipping machines with the human brain’s self-learning abilities. If AI could realize the simulation of the brain through machines, it meant matter could simulate the mind. The simulation is matter’s representation of the mind, undermining the mysteriousness of the human spirit, and therefore was an epistemological breakthrough.

The second, and key issue, was the subject-object relationship. Scholars Cao Boyan and Zhou Wenbin noted that due to the development of production and scientific technology represented by AI, artificial cognitive organs are no longer restricted to the scope of tools for cognition, and might become subjects of cognition; nonetheless, the artificial subjects have no social and class attributes. Their argument created quite a stir in academia; it was so impactful that Tong Tianxiang and Wang Haishan each wrote a paper to debate the outcome. In May, June, and November of 1981, the Philosophical Research editorial board convened three seminars on Marxist epistemology and modern natural sciences in Shanghai, Dalian, and Beijing, successively. The explorations and disputes over artificial cognitive subjects were intense.

The third theme was a classical issue: the relationship between humans and machines. “Machine” is the technical form of AI, but this differs from traditional machines conceptually. “Artificial” indicates machine attributes, while “intelligence” was presupposed to be identical with, or similar to, human intelligence. In 18th-century philosophy, the claim that “humans are machine” invited heated debates, yet in contemporary philosophy, “machines are human” has likewise generated many discussions. AI philosophy can reverse the subject and object in the human-machine relationship, in other words, from “humans are machine” to “machines are human.” Once the proposition “machines are human” is tenable, the relationship between humans and machines is no longer only between humans and technical instruments, or between humans and technical subjects, but also between humans and technical others.

Social turn of research in the 1990s

Entering the 1990s, the Chinese philosophy of AI community’s heated research dynamics cooled down a bit. A social turn occurred in their research focus, as academics cast their sights on AI and society, examining AI in the context of the changes of the times, the social impact of AI, AI and social production, and so forth.

Tong Tianxiang spearheaded this social turn. In his opinion, AI’s development would center on intelligence to usher in a new generation of productive forces, which would take the major form of a “human-machine composite intelligent system” and have the two elements of “high intelligence” and “high-intelligence machines.” New productive forces would bring about new relations of production, and eventually create a new social form—intelligent society. Just as the agricultural society depended on biological resources, and the industrial society regarded technology as the main resource, intelligence would be predominant in the intelligent society. To embrace the ongoing intelligent revolution and the upcoming age of intelligence, the philosophy of AI should also think ahead, Tong suggested.

As one of the founders of the AI domain in China, renowned Chinese mathematician and computer scientist Ma Xiwen is also a great figure remembered for his contributions to the philosophy of AI. He participated in translating and reviewing American philosopher Hubert L. Dreyfus’s What Computers Can’t Do and American scholar of cognitive science Douglas R. Hofstadter’s Gödel, Escher, Bach: An Eternal Golden Braid. Both the English originals and their translations are classic. Ma articulated, “the computer is the steam engine [in the Industrial Revolution],” and called for greater attention to changes taking place as human society enters the information age. “The computer shouldn’t and won’t be the last intelligent machine,” he said, adding that it still has a long way to go.

Vertically, the social turn has been inherited and developed by academia today, which is accustomed to the philosophical study of AI from such perspectives as ethics of science and technology, and social governance.

Originating in the 1960s-70s, thriving in the 1980s, and turning towards society in the 1990s, the philosophy of AI in 20th-century China laid the foundation for the field to develop in more breadth and greater depth in the 21st century.

Lei Huanjie is from the Institute of Philosophy at CASS.