Stines 1

Elliott Stines

Instructor: Malcolm Campbell

English 1103

October 31, 2019

Creative Computers: Science Fiction or Reality?

Imagine you are visiting an art exhibit. Everything seems typical. The event has public

attendance, press coverage, and features works produced by an artist over several years.

However, this exhibit has one key difference: the artist. The works were produced by a robot, not

a human painter. Named the “Painting Fool,” the robot utilizes artificial intelligence (AI) to

produce numerous pieces of artwork. While this may sound like a technology of the future, this

exhibit actually happened at the Galerie Oberkampf in 2013, and the technology has only

continued to improve in recent years with advancements in artificial intelligence technology.

Very engaging intro i really liked it

After I read an ​MIT Technology Review​ article describing this event, I was extremely

curious, the idea that computers could produce works that are associated with creativity

fascinated me. The technology is referred to as computational creativity or artificial creativity,

and simply put, is the utilization of computers to emulate human creativity. As someone with

interest and experience in both programming and graphic design, I found the topic intriguing

from both a technical perspective and an artistic perspective. Creativity is considered by most as

an almost uniquely human trait. No other species on Earth comes close to the level of creativity
 Stines 2

as humans display. However, over the last decade, we have acquired the ability to do incredible

things with computers. With advancements in AI technology, computers can now recognize

faces, translate between languages, take calls for you, beat players at the world’s most

complicated board games, and much more. For the first time in history, our ability to be creative

may not be unrivaled.

I wanted to learn more so I began researching the topic and was surprised to learn that

“creative” AI systems are being utilized to produce not just paintings, but also novels, music,

movie trailers, and more works that we traditionally associate with human creativity. For

example, IBM, one of the lead companies pursuing computational creativity technology, has

created a system that when given one ingredient to start with, will generate a recipe using an

algorithm to combine ingredients that are not typically paired together, based on factors such as

flavor and chemical composition. When viewed by chefs the recipes were often rated as being

both tasty and creative, often pairing ingredients that humans would typically never consider.

Was this machine creative? You’re likely to find a variety of answers to the question.

I was shocked at the wide range of professions contributing to the conversation around

computational creativity. The topic is discussed not just amongst AI researchers, but also

neuroscientists, philosophers, psychologists, artists, and others, with much debate on the extent

of the ability of computers to be creative. Additionally, some artists are vocally against the

technology, stating that a work created by a computer is not truly creative and that such works

lack the human element of art.

The question of the achievability of computational creativity is a nuanced one. Defining

creativity is a difficult task, no single definition seems to offer the complete scope of the concept.
 Stines 3

However, experts tend to agree that we perceive creativity when shown something that is a

combination of novelty and usefulness. The usual argument against computers being creative is

that they are only capable of doing what they are told to do. Humans are the ones that have

programmed them, they are just following the instructions and carrying out tasks accordingly.

However, advancements in AI and machine learning have blurred this line, while operating

under an algorithm, the AI system is able to “learn” through trial and error, similar to the way

humans learn. Through this learning, the system can produce results that it was not explicitly

programmed to produce. One of the greatest challenges in the development of creative artificial

intelligence systems is proving that the software is creative, rather than just an extension of the

developer’s creativity. A computer program that relies on human judgments while it creates can

justifiably be interpreted as an extension of the operator’s creativity. The requirements for

declaring a system creative is where the majority of the debate lies around computational

creativity.

One expert on the subject, Kyle E. Jennings, part of the University of California,

Berkley’s Department of Psychology, specializes in research regarding creativity and the creative

process and has written multiple research articles on the subject. He argues that “creative

autonomy,” a concept that exists when a system is capable of “both evaluating its creations

without outside judgment, and of changing its standards for evaluating its work in a way that is

not random,” is a necessary condition for making the argument that a system is truly creative.

Meaning that a system must be able to evaluate its own work, and evolve its standards for

evaluation in a non-random way, in order for it to be considered creative. Rather than requiring

that the system be completely sealed to avoid human influence, Jennings suggests that
 Stines 4

developing creative autonomy is argued to be more plausible if the system is “intimately

embedded in a broader society of other creators and critics, whether they be human or other

artificial intelligence systems.” However, he states that such a system has not yet been

demonstrated.

Another take on the conditions required for a system to be creative comes from the

creator of the robot initially mentioned in the introduction. Simon Colton, a professor of

computational creativity at Goldsmiths College, London. He suggests that if programs are to

count as creative, they must behave in ways that are“skillful,” “appreciative,” and ​“imaginative.”

The imaginative component of these requirements is demonstrated in another project utilizing

artificial intelligence that captures one aspect of human imagination: the ability to see one thing

as something else. Google’s Brain AI research team developed a system that when given an

image, could detect patterns and shapes, and “hallucinate” images on top of them. The ability to

see an object as something else has been utilized by artists for centuries. Almost everyone has

experienced this phenomenon in simple ways like looking at a cloud and visualizing an animal or

object. Additionally, this capacity might have been one of the triggers for some of the earliest

known art, prehistoric cave art. With prehistoric art utilizing natural features like a pebble in the

wall that looks like an eye, for example. (Gayford np) This suggests that this type of imagination

may have been crucial to the foundation of human creativity and may be just as critical in the

foundations of computational creativity systems.

The visual is interesting I like it

A demonstration of the “hallucination” effect from Google’s Brain AI research.

However, not all experts agree that true creativity is a goal that AI will be able to achieve

utilizing current technology. Some argue that the current structure of neural networks, the

primary technology behind these AI systems, limits the possibility that they will ever have true

artificial creativity. Sociology professor Anton Oleinik argues that “Neural networks are machine

learning algorithms composed of layers of calculations that excel at ingesting vast amounts of

data and finding every pattern within them. They fundamentally rely on statistical

regression–which means that while they’re good at identifying patterns, they fail miserably to

anticipate when a pattern will change, let alone connect one pattern to an unrelated pattern, a

crucial ingredient in creativity.” He goes on to argue that because all patterns appear to be

meaningful to an algorithm based purely on how prevalent they are, “neural networks fail to

distinguish between which patterns are meaningful and which aren’t.” Computers may come up

with novel ideas, but they may not be valuable ideas. However, neural networks can still be

excellent mimickers of creativity but are not capable of meaningfully changing and rewriting the

imposed rules.
 Stines 6

I find this last take the most persuasive, while computational creativity is an impressive

technology in its current state, machines are not creative in the same sense that humans are

creative. Creativity involves a mix of social, emotional, historical, and other factors that are

extremely difficult for a computer to emulate with current technologies. However, this does not

mean that machines have no part to play with respect to creativity. These creative AI systems

offer artists a valuable tool, a new creative collaborator. When put together, AI and humans can

achieve impressive things. They can become the best game players in the world, transfer styles

of famous paintings to any image, and find solutions to problems we thought impossible to solve.

Maybe try to seperate the last paragraph into 2, discuss the topic then have a proper

conclusion, i feel like it just ends pretty abruptly.

Works Cited

Gayford, Martin. “Robot Art Raises Questions about Human Creativity.” ​MIT Technology

Review, ​15 Feb. 2016,

www.technologyreview.com/s/600762/robot-art-raises-questions-about-human-creativity​.

Accessed 20 Oct. 2019.

Haynes, Suyin. “This Robot Artist Just Became the First to Stage a Solo Exhibition. What Does

That Say About Creativity?” ​TIME​, 17 June 2019,

www.time.com/5607191/robot-artist-ai-da-artificial-intelligence-creativity​. Accessed 16

Oct. 2019.

Jennings, Kyle. “Developing Creativity: Artificial Barriers in Artificial Intelligence.” ​Minds and

Machines​, vol. 20, no. 4, 2010, pp. 489-501, ​www.doi.org/10.1007/s11023-010-9206-y​.

Accessed 16 Oct. 2019.

Kulpati, Sarvasv. “Can AI Be Creative? A Comprehensive Look at the State of Computers and

Creativity” ​Towards Data Science​, 28 July 2018,

www.towardsdatascience.com/can-ai-be-creative-2f84c5c73dca​. Accessed 28 Oct. 2019.

Schwab, Katharine. “3 Reasons Why AI Will Never Match Human Creativity” ​Fast Company​,

25 Mar. 2019,

www.fastcompany.com/90339590/3-reasons-why-ai-will-never-match-human-creativity​.

Accessed 18 Oct. 2019.