Staff writer Amy Meng attended a talk featuring science-fiction writer Ted Chiang hosted by the Asian American Initiative, in which he discussed the ethicality, legality, and philosophy behind AI-generated content in the artistic world. Here are some of her takeaways. 

On the evening of Wednesday, April 22, I attended an Asian American Initiative lecture event, “The Incompatibilities Between Generative AI and Art,” which featured well-known American science fiction writer Ted Chiang, followed by a discussion with Columbia Professor of English and Comparative Literature Denise Cruz. Upon receiving the initial email about the event, I knew I had to go, having read his book Stories of Your Life and Others. Additionally, as a visual artist and creative writer, I was intrigued to see what Chiang may have to say about the rise of AI in artistic spaces. Here are some of my key takeaways and favorite quotes from Chiang’s talk. 

“Art is a concentrated form of intention.”—The difference between a red canvas and red rectangle. 

Multiple philosophers and concepts were mentioned throughout Chiang’s talk, reinforcing his arguments. One of the first philosophers he mentioned was Arthur Danto, whose book, The Transfiguration of the Commonplace, establishes a clear distinction between a canvas with red paint and a red rectangle. While the painter’s message might only be legible from the piece’s title, this is exactly what separates it from being a mere object. “What I think Danto to be saying is that careful deliberation is enough to make something art,” Chiang said. This concept is especially important in a world of contemporary art forms; what makes the artwork might not be the appearance itself, but the intention behind its creation. 

Chiang also established the connection between the artist’s intention and the audience’s appreciation. “There may be many ways to create a focused appreciation in a reader or a viewer…but I don’t believe that generative AI can revoke a similar focus of appreciation,” Chiang said. Reflecting upon his claim, I also found myself positing the fact that the greatest appreciation for art comes from the artist’s message. There might be a person behind the screen, inputting text blurbs for the AI machine, but the reality is that AI-generated art lacks the emotions and experiences that a human does. So why should AI-generated art deserve anyone’s appreciation? 

“The value of art is not solely inherent to the object.”

This is a line that especially stood out to me. As an artist myself, I also found this statement to be deeply resonating. The artistic value of the final artwork, the “object,” goes beyond its visual characteristics. The entire artistic process can also be viewed as art; the initial brainstorming, repetitive brushstrokes against the canvas, and emotional rollercoasters that go through the artist’s mind as they create. This is all art. 

“In the modern media landscape, your attention is perhaps the most valuable thing you possess.”— The time it takes to create art should be greater than the time it takes to consume art. 

In bringing up a scenario in which someone had produced a short story using a generative AI tool tailored to creative writing, Chiang explained the difference in the time it takes an artist to create their art and the time it takes the audience to consume the art. It may take us a few hours to read a book, but months and even years for a writer to complete it. A music video may only last three minutes, but so much more time was spent by the singer, songwriter, and music producers to ensure the final product. Chiang said that he would “spend more time reading a comment on Facebook than anything generated by a large language model because assuming that Facebook comment was written by a human, more time went into its composition, so it deserves more of my attention.” The Facebook user’s comment was crafted from a backdrop of opinions and emotions; they took time out of their day to write this comment. What AI’s product was crafted from is a massive pile of scattered information that comes from all corners of the internet. There is minimal effort from the person behind the screen.  

Thinking about generative AI as a car chop shop.

I found this comparison that Chiang made to be extremely effective. He compares the way that generative AI functions to that of chop shops, which disassemble cars into parts to be redistributed among many other cars. There is much disconnection between the cars that come into the shop and the cars that leave the shop as repaired or crafted products. In both of these scenarios, “the product being sold in untraceability,” as Chiang puts it. Similar to chop shops, generative AI machines are taking snippets of information from countless sources, yet oftentimes not offering accurate citations. Take this as a cautionary tale for your future classes, if you haven’t yet. Current generative AI models are not the best at quotations or citations. Sometimes, the quotations simply do not exist. 

Additionally, even if the original source is credited by AI, AI companies may not even have been permitted to use it. Chiang compares this to money laundering. “You might accept a fee if the Chop Shop has already chopped up your car and the alternative is nothing, but that is different than agreeing that the price is fair,” Chiang said. “As long as the Chop Shop remains a profitable business, it is doing so by taking money that belongs to you.” AI-generated content brings up both ethical and financial concerns. 

“The more that you use generative AI to produce fiction or paintings, the more you erode your own ability to recognize what self-expression entails.” 

Chiang also delves into the fact that entering a prompt does not fall into the same category of endeavor as writing a book or creating a painting. Indeed, these are vastly different processes. The former takes the easiest shortcut in producing something that may seem like “art.” On the other hand, the latter is a product of careful deliberation and targeted messaging. In such a digitized world in which everyone, especially younger generations, is becoming increasingly dependent on generative AI, it is important to distinguish between one’s own artistic expression and an emotionless product of artificial intelligence. 

AI as a brainstorming tool vs. AI as a text generator.

There is a major difference between using AI as a source of inspiration and using AI as a text generator, as Chiang argues. He prefaces this argument by establishing that “machine learning could be useful, as long as it does not reduce the artist’s control over the final product.” When the artist’s control over their own creation is impeded or even taken over, can they really call the final product their own? From my perspective, the way AI may impede artistic freedom depends on the scope in which it is used, as well as the timeline. However, even if one is solely using AI as a brainstorming tool in their writing assignment, for instance, they will start to develop this dependence on the machine. Unfortunately, the reality is that many regular generative AI users are dependent on these machines to create for them. 

A quote from philosopher Hannah Arendt: “The trouble with modern theories of behaviorism is not that they are wrong, but that they can become true.”

One of the other philosophers that Chiang mentioned was Arendt. His interpretation of this quote of hers is that “if we build a world that treats humans as nothing more than stimulus response pairings, it will erode our capacity to be anything more than bundles of stimulus response pairings.” This is a world that most of us, I hypothesize, do not desire. Humans are indeed so much more than “bundles of stimulus response pairings.” We have experiences, emotions, and perspectives that mold us into the people we are today, that mold our art into what it is. 

Ted Chiang ended his talk with inspiration: “Human beings don’t have immaterial souls, but we are capable of making meaning, and art is one of the ways that we do that. So, I encourage you to go forth and create something that is meaningful to you, because by doing so, you have a good chance of creating something that is meaningful to someone else. And that is what we are here on Earth to do.” Art is all about personal creation, intention, and appreciation. The audience is just as important as the creator, in many cases, as it creates room for artistic expression and ideological dialogue. These are all examples of what generative AI will never be able to accomplish. 

Image via Author