Shi Zheng: Is the Return to Normalcy an Illusion?
Written by Zhang Anya
Today, January 31, 2023, marks the end of the Spring Festival holiday for most people, who are returning to their work routines; yesterday, the Health Commission announced that the national epidemic has entered a low prevalence level, officially making memories related to the pandemic a part of history. It seems that various signs are indicating that life is returning to "normalcy." However, the "abnormal" still lurks in the corners. For instance, on the first day back at work, one might physically sit in the office while mentally still catching up on sleep; or one might instinctively reach for a mask in crowded settings; or hesitate to make long-term plans, instead reminding oneself to cherish the present moment… A calm, predictable life still coexists with the "abnormal."
Under the theme "The Abnormal Everyday," we have invited the artist Shi Zheng to create a special artistic cover for this issue. The artwork depicts a routine moment involving a machine, incorporating two of the artist's classic works.
A clever nested "picture within a picture" effect is achieved with the placement of the Life and Arts magazine on the desk. Reality and illusion intertwine within the constructed space, each element subtly infused with metaphors for reality.
Shi Zheng
Like many of the works of this artist, born in the 1990s, there is a distinct cyberpunk perspective, aiming to create a pure perceptual experience through the use of media imagery.
· The Illusory Office ·
The desk in the image might seem familiar to you— the arrangement of the seats, books casually placed on the desk, and two yellow sticky notes on the corner of the screen—all subtly suggesting an extremely ordinary moment in daily life. However, the excessively uniform and smooth color tones impart a hint of eeriness, casting doubt on whether this scene truly exists.
The cover of collabration of Shi Zheng and Life and Arts magazine (detail)
In fact, this is a 3D computer-simulated rendering of Shi Zheng's own office. What appears as a simple, ordinary visual portrayal becomes redefined through modeling, creating a mysterious atmosphere filled with underlying currents. Similar to his work From the Film Set of <Alphaville>, Shi Zheng draws inspiration from Jean-Luc Godard's Alphaville. It seems as if the owner of the desk had just left a moment ago, yet the clues given by the blinking lights of the operating hard drive suggest that the machine is always present.
Shi Zheng, From the Film Set of <Alphaville>, 2020
Film Alphaville
Alphaville tells the story of a machine's awakening, where the extreme technocratic supercomputer "Alpha 60" controls the entire city. Shi Zheng attempts to evoke a feeling of "machines present but human whereabouts unknown." It might be challenging to determine which is the real "film set": the visual world constructed by the image content or the lines of code executed by this unnamed machine.
· The Unreadable Newspaper ·
On another note, if you closely observe the screens, wall decals, and even the cover of magazines in this virtual space, you will find that they all blur into the front page of The New York Times,"metaphysically merging the "abnormal" with the "everyday."
Shi Zheng x Life and Arts co-create cover details
These indistinct images are, in fact, generated through an AI algorithm, a methodology central to Shi Zheng's hallmark series Frosty Morning. The New York Times, symbolizing the mundane, loses its readability here, though its impact lingers in headlines, layout, and other formal elements, guiding contemplations on time and reality. In 2018, Shi Zheng collected over 2000 images from the front pages of The New York Times over the past six years, organized them chronologically, and trained a neural network to iteratively learn and produce an endlessly generating digital multimedia depiction of these newspaper images. This collection, with a sense of predicting the future, comprises raw and more complete versions of the images that have once graced the pages of The New York Times, symbolizing the hidden realms of memory and history.
Shi Zheng, Frosty, Morning, 2019
"Time is like a flame in which we burn," Shi Zheng contemplates that this poetic expression aptly describes the newspaper images generated by AI. The burning encapsulates time and space of the past—an irreversible series of events—as well as our ongoing present, converging with each imminently flowing second into the future. Simultaneously, news media, typically associated with objectivity and facts, reveal their poetic dimension in artistic expression.
The New York Times holds a pivotal position in the history of art. It has been reported how this publication, integrating geopolitics and social issues, human stories, and trending patterns, and organizing cultural progressions like a chronicle, has become an "inspirational muse" for artists. Artists such as Fred Tomaselli and Shibuya Sho fuse art with news by reimagining the pages of The New York Times, thereby blending artistic expression with journalistic elements.
Fred Tomaselli，The Times
Dave Mckenzie, Yesterday’s Newspaper
Shibuya Sho, From the series Sunrise Through a Small Window
Not only does The New York Times serve as an example, but newspapers, as a prevalent form of mass media, document real events occurring around the world, thereby becoming symbols of public opinion and the zeitgeist. When the paper that carries the news no longer possesses immediacy, it transforms into a medium for artists to explore the world and reconstruct the zeitgeist. The "rational" nature of news and the "emotional" nature of art represent two seemingly disparate currents of reality that come together to complement and fulfill each area.
It is believed that only a professional perspective from media studies can truly spark dialogue with this kind of artwork. Communication scholar and renowned journalist Fang Kecheng observes and interprets the world from a media perspective. The Life and Arts magazine invites Fang Kecheng to engage in a conversation with Shi Zheng about algorithms, news, and creative processes. Through this dialogue, by examining the artworks themselves, we can access the logic behind Shi Zheng's creations and pose questions that are more insightful: What is the significance of art and viewing in an era where artificial intelligence technology rapidly evolves? Amidst the flow of thoughts, the "tech-engaged future" described by Shi Zheng becomes increasingly vivid.
「INTERVIEW}
Fang Kecheng: Fang
Shi Zheng: Shi
SECTION: A
How to Redefine the Power of "News"?
Fang: In your most recent version, you have placed the newspaper in an everyday setting. Why did you choose this approach?
Shi: I have been continuously collecting images of newspapers that have been discarded in everyday contexts, somewhat akin to stumbling upon a scene unintentionally. For instance, one piece features a New York Times headline commemorating the 50th anniversary of the moon landing, with an illustration derived from a photo taken by an astronaut of their own foot and footprint on the moon. When I looked at it from an overhead perspective, I felt as though I was being drawn into that scene. This viewpoint transcends time and the spatial relationship between us and the newspaper.
Shi Zheng, Frosty Morning (Daily), 2021
Shi Zheng took an photo of The New York Time been put on the ground, inspired him to create his new work Frosty Morning (Daily)
Fang: Newspapers constitute an important part of everyday life, yet the content they contain often pertains to the non-everyday, possessing what journalism refers to as newsworthy value. From your point of view, I sense another tension in this work, namely, the tension between the ordinary and the extraordinary.
Shi: In fact, I recall that in 2019, there was substantial discussion around fake news, during which some individuals printed counterfeit newspapers. Because these forgeries mimicked the conventional perspective from which newspapers are viewed, many people were deceived into thinking they were genuine. Including in Frosty Morning, I am not particularly concerned with the actual content displayed; rather, I am interested in witnessing its transformations in fleeting moments. Since it has adopted the layout of a newspaper, it continuously maintains an appearance very similar to that of a newspaper, giving an impression as if it were indeed the real thing.
施政，《BINE》，2019
In the derivative work Frosty Morning, Shi Zheng removes all images and text from the newspaper, leaving only the outlines between columns. Machine learning is then used to generate images of continuously spreading intersecting lines, to which random colors are assigned. Even though the text and images can no longer be read, Shi Zheng believes that these flowing lines, under the layout of The New York Times, sufficiently represent those social processes that have occurred and appeared.
Fang: This brings me to an interesting point. Over 10 or 20 years ago, it was commonly said that newspapers were dying or already dead. Yet, both you and other artists continue to choose newspapers as a medium or form. This suggests that newspapers indeed have a vital presence. From where does this vitality or allure originate?
Shi: Perhaps it stems from their material-based nature which, compared to the constantly changing information on websites, makes physical newspapers appear more trustworthy. Additionally, the front page of a newspaper has a certain definiteness and rhythm, appearing anew each day, unlike website headlines that may change every few hours or even minutes. On algorithm-driven platforms, everyone sees something different.
Fang: Why did you not use social media or The New York Times website as your material?
Shi: The answer is quite simple. I can quickly obtain PDFs of the daily newspaper layout from its website. I have considered using other newspapers as well, but found the collection process to be very lengthy and cumbersome. The New York Times makes its content so easily accessible for experimentation and rapid results.
Fang: Indeed, although these newspapers are fundamentally commercial entities, they provide materials that serve, in some aspects, as public resources. When discussing the societal functions of the media, we seldom mention its role in offering materials for artists and other types of creators, which is quite fascinating, including the resources it provides for scholars. What kind of media sources do people tend to choose? Naturally, those from which materials are more readily obtainable. Therefore, the role of media and news is not limited to being consumed by the public the day after publication; it also serves as a significant repository, an archive here.
Section B:
Are humans no longer the sole observers in the world?
Fang: Your work utilizes machine learning to generate images. Have you sought to convey that the future could be repetitive, or perhaps that the future is predictable?
Shi: I contend that the future is not merely a concept, but the outcome of technological engagement. The foundation of Frosty Morning lies in an unsupervised learning technique based on Generative Adversarial Networks (GANs), structured primarily around two models: a "generator" and a "discriminator." The "generator" produces images that mimic reality based on real samples, while the "discriminator" assesses whether the given images appear "natural/real." What we observe can be envisioned as a strategic contest between two neural networks, continuously challenging the notion of "authenticity." Interestingly, although we humans may not be able to interpret these images, the process can be regarded as a form of 'reading' by machines as they are trained to process these image datasets. It is somewhat akin to the idea that the visual perception we currently share is in fact a shared experience with machines.
In my subsequent work, Free Fall, I utilized a facial recognition algorithm to analyze whether falling face masks resembled human faces. Essentially, this piece aims to convey that in the world displayed on our screens, humans are no longer the sole observers; machines are also watching us.
Shi Zheng, Free Fall, 2020
Fang: As a verb, 'observing' inherently implies subjectivity within its subject, and the subject of observation is often the one with power, also suggesting a relationship of authority. When you mention that so much visual material has been collected, and it is not humans but machines that are observing, do you think that machines can now be considered as having a degree of subjectivity, and that the power dynamics between humans and machines are also changing?
Shi: I believe there is indeed such a trend where both humans and machines are evolving. This is precisely why I think this artwork might make some viewers uncomfortable because they are unable to interpret it; they find it very painful. They cannot clearly see this constantly changing generation of text or images. For a digital work, is the video itself the artwork? Rather, the video we see and the neural network that has been trained with these images could be considered the artwork itself.
Fang: Indeed, the point you make is very intriguing. When you use machine learning as a tool for creation, have you ever wondered who truly authorizes the artwork?
Shi: I think this question has always existed, dating back to the concept art era, where readymades were considered art, posing a similar dilemma. In truth, I cannot provide a definitive answer. I feel that technology assists me in completing a work. If I were to become deeply engrossed in this issue, as a creator, I would find myself at a loss on how to proceed, feeling constrained.
Fang: With the advancement of artificial intelligence, do you think it's possible that, after learning from your previous works, these machines could independently create another piece attributed to Shi Zheng?
Shi: Absolutely, as long as I produce enough works. If someday I can simply input a command like "a video work like Shizheng" and receive an output, I would indeed feel a sense of satisfaction. At that time, I could also engage in other activities, which highlights a characteristic of human subjectivity.
Photographer Stephen Shore shared on Instagram images generated by the AI algorithm DALL-E after inputting "a work like Stephen Shore." Shi Zheng found this particularly intriguing.
Fang: So overall, your attitude towards artificial intelligence is quite measured. How do you view the public's romanticized expectations or excessive fears concerning artificial intelligence?
Shi: In fact, from the perspective of image learning, the entities machine learning creates are trained by humans. For instance, in my work "Free Fall," the label "human" that it identifies is actually assigned by the facial recognition algorithm I programmed. I believe this also constitutes a form of deception for the viewers. The logic keeps repeating; I am deceiving the machine, and the machine is deceiving people.
Fang: Would it be accurate to summarize that, given the underlying principles of AI today, machines still lack a creative existence?
Shi: In fact, some images already possess a certain degree of creativity, though they are not generated out of thin air. For instance, the scholar Lev Manovich, who conducts research in neural aesthetics, has been experimenting with text-to-image generation. This technology gained significant popularity earlier this year, but I have subjectively maintained a distance. For me, assuming I were to use this technology, I would let it serve as a source of inspiration for my work. When I cannot find a similar image, this algorithm could help me generate one, and I would then create something of my own based on it. I typically wouldn’t consider using it directly as a piece of art, although some artists have already exhibited such generated works. I believe the openness of art allows everyone to engage in various activities, but the subjective choices of artists actually vary from one individual to another. Returning to your question, I think that, in the future, this matter actually involves technological participation, and the so-called prediction of "the future" is actually feedback from past times.
Artwork generated using the Midjourney software by Lev Manovich.
SECTION: C
"Artificial Intelligence" is still not truly “intelligent"?
Shi: Many of my works are related to simulated scenarios. I once created a photograph using a computer to model a corner of my own studio. In the image, there is a computer with its screen turned off, but the light on its hard drive is illuminated. I titled it From the Film Set of <Alphaville>, referencing a film called Alphaville, which depicts a city controlled by computer simulations. This work transforms the context of viewing, placing us in an environment akin to a movie set.
Fang: Are you interested in the concept of simulation in relation to the so-called theory proposed by some, suggesting that we live in a simulation?
Shi: I perceive my simulations as a form of rehearsal for reality. In actuality, I am using simulations to reproduce natural phenomena. For instance, I recently created a scene of a melting iceberg, which was inspired by footage of an iceberg collapsing in Greenland captured by someone else. I reran the process using a computer, resulting in an image that offers an alternative viewing perspective, where time can be extensively elongated. The gradual dissolution of the iceberg, which is inherently a very slow process, generates a sense of estrangement. This simulation is also a way of viewing that I desire, as it enables me to see certain images that encompass content I wish to see, including events that might ordinarily seem unlikely to occur.
Shi Zheng, Melting in Time, 2022
Fang: So, it is essentially your own way of constructing a world, although this may not be a complete world but rather an extraction of some of its elements.
Shi: Indeed. Actually, the piece Frosty Morning also seems like a flat window through which one views a world—only it is a newspaper. The outside part of this world, although not directly displayed to others, is nonetheless presented through the newspaper as a visage of the world, or perhaps more like a portrait.
Fang: From a distance, the machine-generated layouts of The New York Times seem quite convincing, but upon closer inspection, they are revealed to be meaningless pixels. In Free Fall, the machine also earnestly scores face masks; does this suggest that artificial intelligence is still not very intelligent?
Shi: The title of my exhibition, "Slothful Machine," actually hints to the audience that these ostensibly high-tech algorithms are particularly unintelligent. They can generate text that you cannot understand, and they might mistake a face mask for a human face. More so, they provide a space for imagination, proposing a hypothesis that perhaps these machines are in a state of indolence, slowly altering the generation of these newspapers.
Fang: The term "Slothful" is quite intriguing. I consider it to denote a significantly limited existence. When AI is engaged in generating newspapers or recognizing faces, it may appear diligent, but beyond these tasks, it seems “unwilling” to do more, which is what is meant by indolence.
Shi: It actually represents a reduction in agency. However, I believe that indolence more closely signifies a state of potential energy, akin to being at the lowest point of a curve, poised to rise. I trust that this will be the trend in the development of artificial intelligence and machinery.
Fang: So perhaps in five years your exhibition will be titled "Unslothful Machines."
