AI Narratives in Contemporary Chinese Science Fictions

A seminar co-hosted by Berggruen Research Center at Peking University and Leverhulme Centre for the Future of Intelligence at the University of Cambridge, titled “AI Narratives in Contemporary Chinese Science Fiction,” was livestreamed on Bilibili on November 17, 2020. Sci-fi has existed in some form for at least two centuries. Since its modern inception with the publication of Mary Shelley’s Frankenstein (1818), the genre has evolved far beyond mere entertainment to become a substantial influence on human culture and progress. By zooming in on contemporary Chinese AI-based sci-fi, the seminar offered a clear and pertinent introduction of how AI is depicted in one of the most influential global literary subcultures over the last four decades. Sci-fi scholars Wu Yan (吴岩) and Sanfeng (三丰)  delivered opening keynotes, while the seminar’s second half featured a roundtable discussion of four sci-fi writers: Baoshu (宝树), Chen Qiufan (陈楸帆), Fei Dao (飞氘), and Xia Jia (夏笳). They shared their insights on several topics related to AI and science fiction. Hao Jingfang (郝景芳), novelist and 2019-2020 Berggruen China Center Fellow, served as moderator.

Chinese Science Fiction in the Early Period of Reform and Opening Up (1978-1983)
Early references to artificial intelligence can actually be found in many Chinese science fiction works dating back to before the period Reform and Opening Up that began in the late 1970s. In his remarks, Wu Yan of the Research Center for Science and Human Imagination at Southern University of Science and Technology discussed a number of representative stories. According to Wu, the boundaries in these stories between computers (or “electronic brains,” a literal translation of a common nickname for computers in the Chinese language), robots, and artificial intelligence were quite blurred, while all tended to be depicted as educational and obedient servants to humankind. Notably, the term “artificial intelligence” even appears as early as in these pre-reform stories.

This resurgence of Chinese science fiction began in 1976, and the years between 1978 and 1983 saw a number of “quasi-AI” stories gain wide acceptance. During this period, some authors began deeper explorations of AI. Xiao Jianheng (萧建亨), through his examination of algorithms, machine learning, and situational awareness, identifies this time as a remarkable moment of enlightenment in the Chinese literary depictions of intelligent robots. Concurrently, Wei Yahua (魏雅华), whose influential works were translated and exported soon after their publication in China, pondered moral reconstruction at the social and cultural level. We can roughly describe the evolution of Chinese science fiction during this early period of the Reform and Opening Up as an evolution from robots to algorithms to autonomous learning in ecological and cultural contexts, and a shift from situational awareness to the establishment of moral paradigms. The public at that time, who initially saw robots as little more than obedient servants, began to grow concerned with the moral codes governing artificial intelligence, and later generations began to worry about the now-well known issues of “intentional camouflage” and the potential for rebellious AI.

Since 1984 AI storytelling has taken off in China, more recently growing in both quantity and variety. AI fiction today often tells stories of competition between humans and post-human species and ensuing technical or ethical conflicts. Wu believes that recent fiction owes much to its predecessors from the early reform years; these, along with the stories written in the “transitional period” between 1984 and 2000, deserve more thorough study.

Algorithm of the Soul: AI Narratives in Recent Chinese Science Fiction
Recent Chinese AI sci-fi highlights the ongoing development of AI technology and the unique position it holds in contemporary life. Sanfeng, a visiting researcher at the Research Center for Science and Human Imagination at Southern University of Science and Technology describes two important features of the last decade’s groundbreaking sci-fi stories.

1. Symbolicism, also known as logicism or computationalism, one of the two major approaches to AI, advocates using axiomatic logic systems to build a set of guiding rules for an intelligent agent. It has gradually lost its dominant position to the other school of thought: connectionism. A form of bionics, connectionism suggests that we should design AI to emulate human neurons, creating systems similar to neural networks in the brain. Connectionism underlies well-known contemporary concepts such as big data, neural networks, and deep learning. A pioneering example of “connectionist” sci-fi can be found in Ted Chiang’s The Lifecycle of Software Objects.

The shift towards connectionism has been keenly grasped by Chinese sci-fi writers. Xia Jia (Let’s have a talk), Hao Jingfang (Where Are You, 你在哪里), Chen Qiufan aka Stanley Chan (Image Builder, 造像者), and Mu Ming (慕明) (Forging Dreams, 铸梦), among others, have all tried to trace the origins of the robotic soul or consciousness; algorithms based on neural networks have now emerged as a popular focal point in their work. According to Sanfeng, the credibility of AI’s characterization in sci-fi stories increases when it is inspired or corroborated by algorithm theories.

2. How are we to interpret tension between AI and humans in sci-fi stories? To sharpen this question, Sanfeng breaks sci-fi down into three categories: near-future, mid-future, and far-future. Faced with an increasingly AI-filled reality, Sanfeng says, sci-fi writers should focus on the real world—on the conflicts, emotional or ethical, caused by the ubiquitous presence of AI today and in our near future.

Readers today expect sci-fi writers to understand, depict, and respond to AI’s application in the real world and emotional impact on ordinary people in recognizable ways. Difficult as it is, a number of young authors have lived up to these standards. It is hard to exaggerate the influence of Black Mirror on recent Chinese AI science fiction, such as Into the River (涉江) by Xia Jia, My Girl (妞妞) by Baoshu, The Algorithms for Life (人生算法) by Chen Qiufan, and The Question of Love (爱的问题) by Hao Jingfang, which all depict the near future. That’s what Sanfeng hopes science fiction can do: stay one step ahead of science, so that it can be both a pioneering force and a salve for a fearful public. Sanfeng raises The Stories of Ibis by Hiroshi Yamamoto as an example of sci-fi that portrays a promising future where humans and AI coexist in collaborative and symbiotic relationships.

Is Strong AI Coming?
As we know, the AI “singularity” refers to the moment when the human brain is fully modeled and artificial intelligence outstrips its human progenitor in every way. Not surprisingly, many people are terrified of that version of tomorrow.

As a sci-fi author, Baoshu believes that AI is boundless in its capabilities, and will surpass any limits we can imagine. In other words, there is no way to rule out the possibility of strong AI. Three years ago, AlphaGo taught us a groundbreaking lesson about the might of neural networks and deep learning: algorithms allow non-human entities to sidestep cognitive limitations. Speaking from his writing experience, Baoshu says that strong AI can feel other-worldly and dreadful, and that it’s challenging to imagine something that surpasses the human mind to the degree that it almost resembles divinity. How do human minds function? Does subjective human consciousness emerge from nothing, or is it the effect of neural collaborative processes? Indeed, our self-understanding is quite rudimentary, with many questions still in the stages of preliminary investigation. Given that current AI systems are, for the most part, learning and imitating how humans perceive, conceptualize, and make decisions, the current approach to AI, even if empowered with deep learning and multilayer neural networks, is unlikely to deliver the kind of “strong AI” we imagine. Yet things might be different if AI can somehow be designed in a radically different approach that results in a type of cognition orthogonal to the human mind. Fundamental paradigm shifts would be necessary when talking about whether mutual empathy and understanding among different species is possible. Through these unknowns, Baoshu continues to keep an open mind about a future era of strong AI.

Fei Dao, sci-fi author and associate professor at the Department of Chinese Language and Literature at Tsinghua University, thinks that coexistence and mutual trust could be extremely difficult for agents with such different values as those human and AI may come to have. In a sense, he argues, humans and AI are engaged in a parent-child relationship, since AI’s learned principles and concepts emerge from the data it’s fed by humans. It is humans who are responsible for choosing what values to teach AI. For example, what if we were to teach AI to emulate the human-AI relationship depicted in the movie The Wandering Earth? In the movie, MOSS, an AI system commanding a space station, calculates that it is rational to abandon the humans on Earth who are trying to steer the planet away from an imminently exploding sun, though doing so directly contradicts the values held by the protagonists. According to Fei Dao, the ethical question of how AI will understand these “values” is very interesting.

As the sci-fi writer Chen Qiufan sees it, Fei Dao’s point can be boiled down to a more basic idea: otherness. The debate over human-AI relationships is just the latest manifestation of the idea of otherness and a continuation of the “we-them” discussions we’ve been having since Frankenstein. As Chen sees it, in the present stage of human-AI interaction, we need to bring our agency into play and activate our self-understanding and self-imagination through dialogues with AI. AI is a like a mirror whose refraction allows us to see things invisible from our anthropocentric viewpoint and to approach issues with a deeper perspective.

Xia Jia, sci-fi author and director of the Chinese Literature Department at Xi’an Jiaotong University, offers a literary and cultural perspective on contemporary narratives of “humanoid” robots, pointing out that our imagination of AI as the “other” has been directly influenced by the course of history. In the fictional world of Karel Čapek, the Czech writer and playwright who coined the term “robot” in his play Rossum’s Universal Robots or R.U.R. (1920), just as humans exist under the capitalist system primarily to manufacture physical products, robots are created as a labor force—intelligent but soulless objects. Yet the soul is something unique to humans, and is unable to be restored or sold by the capitalist production system. Strong AI, in parallel, reflects our strong non-physical needs for emotional labor or the capacity to communicate and comprehend as we enter the post-industrial era.

What makes contemporary Chinese science fiction distinct from its western counterparts?
“What makes Chinese science fiction written after the Reform and Opening Up distinct from its western counterparts” was a key topic of the roundtable session proposed by the organizer.

There is no essential “Chineseness,” according to Xia Jia. The differences between China and the West come down to the unique experiences China has had in the process of its modernization. For example, disadvantaged and marginalized groups were active agents in the modern Chinese revolution, using their weaknesses to their advantage to survive and communicate as members of society. Although China was for a long time absent in the imagined worlds of foreign science fiction, Feidao says, the Chinese sci-fi community, initially as readers and then as authors, has captured their concerns about their country’s present and future in their works.

Chen Qiufan admits that science fiction, in its genre characteristics and aesthetic tastes, largely emanates from the West, and a whole generation of Chinese sci-fi authors actually began by imitating their western counterparts. The occasional Chinese elements notwithstanding, these Chinese sci-fi stories, according to Chen, are still rooted in the same fundamental tradition of western science fiction. As he writes, he frequently thinks about how to convey “Chineseness”. Does it suffice to simply apply Chinese expressions, structures, and ideas, or should writers incorporate symbols of Chinese sensibility and appreciation for beauty? Chen discounts these attempts as labored and superficial, and he stipulates that striving for a truly Chinese order of science and technology could bring us a new future worth exploring and imagining. The discourse on Chineseness is only beginning to emerge, he says, and it will continue to be developed by both storytellers and academics.

Baoshu takes a practical view: Chineseness is descriptive instead of normative and ought not to shackle creativity. Yet he recognizes the many differences between Chinese and western science fiction. Since the Reform and Opening Up, attempts by Chinese to understand both the West and their own roots flourished, and narrative motifs in science fiction and local cultural traditions were increasingly combined by Chinese writers in a conscious attempt to identify their cultural roots and reinterpret them against the backdrop of modern technology. Works of both historical and modern Chinese science fiction often embody a grand sense of history or subtle connections to specific Chinese elements, such as lineage, filial piety, and heavenly Tao. Chineseness is thus naturally revealed.

The seminar ended with an emphasis on the common emotional characteristics of humanity and the differences therein. Hao Jingfang maintains that no two persons or communities are the same, and every nation is one of a kind. Therefore, as long as writers are true to their hearts and write from their lived observations and experiences, they will create stories imbued with their national characteristics.

Script by Xu Caidan, undergraduate student at the Chinese Literature Department,
Xi’an Jiaotong University

Edited by Lan Tianmeng, undergraduate student at the School of Philosophy,
Renmin University of China

 

 

 

 


composed by Arswain
machine learning consultation by Anna Tskhovrebov
commissioned by the Berggruen Institute
premiered at the Bradbury Building
downtown Los Angeles
april 22, 2022

Human perception of what sounds “beautiful” is necessarily biased and exclusive. If we are to truly expand our hearing apparatus, and thus our notion of beauty, we must not only shed preconceived sonic associations but also invite creative participation from beings non-human and non-living. We must also begin to cede creative control away from ourselves and toward such beings by encouraging them to exercise their own standards of beauty and collaborate with each other.

Movement I: Alarm Call
‘Alarm Call’ is a long-form composition and sound collage that juxtaposes, combines, and manipulates alarm calls from various human, non-human, and non-living beings. Evolutionary biologists understand the alarm call to be an altruistic behavior between species, who, by warning others of danger, place themselves by instinct in a broader system of belonging. The piece poses the question: how might we hear better to broaden and enhance our sense of belonging in the universe? Might we behave more altruistically if we better heed the calls of – and call out to – non-human beings?

Using granular synthesis, biofeedback, and algorithmic modulation, I fold the human alarm call – the siren – into non-human alarm calls, generating novel “inter-being” sonic collaborations with increasing sophistication and complexity. 

Movement II: A.I.-Truism
A synthesizer piece co-written with an AI in the style of Vangelis’s Blade Runner score, to pay homage to the space of the Bradbury Building.

Movement III: Alarmism
A machine learning model “learns” A.I.Truism and recreates Alarm Call, generating an original fusion of the two.

Movement IV: A.I. Call
A machine learning model “learns” Alarm Call and recreates A.I.Truism, generating an original fusion of the two.


RAVE (IRCAM 2021) https://github.com/acids-ircam/RAVE