Creativity and Technology: Exploring AI Authorship

The article discusses the complex issue of authorship in academia and the art world. It highlights the various practices and controversies surrounding the concept of authorship, including ghost authorship and the contribution of apprentices in creating an artwork. The article also examines the role of AI in authorship and its potential impact on the current understanding of the concept.

Updated on September 29, 2023

ai and authorship

As technology advances, new questions arise regarding authorship. A preprint posted on Research Square speculates that ChatGPT3 met the criteria to be credited as an author on a scientific paper. This raises the question of whether Siri, an AI voice bot, which is also able to answer questions posed by the preprint authors confidently, could also be considered an author.

Authorship is a complex and heavily debated issue in academia. Different levels of credit are given to authors based on their contributions to a paper. Practices such as ghost authorship, and giving credit to individuals who have little or no involvement in a paper, are common but frowned-upon practices. They can compromise the integrity of the publication process.

In stark contrast, the modern interpretation of authorship was almost unnecessary for centuries in the world of art. For instance, according to the National Gallery of Art, Rembrandt had as many as 614 paintings attributed to him. The total number of works attributed to Rembrandt varies depending on a given expert’s opinion. Rembrandt was known to have gone blind during the later stages of his life, and apprentices painted his artwork for him. Many artists are not as prolific as Rembrandt and his apprentices. In comparison, there are ongoing disputes about the approximately 24 paintings attributed to Leonardo Da Vinci.

Authorship in the art world

Successful contemporary artists did not stray too far from old-world authorship. For example, many successful modern artists are credited as the authors of works they did not physically create by themselves, including Louise Bourgeois, Andy Warhol, and Kehinde Wiley.

Andy Warhol was asked by Cavalier Magazine in 1966 whether or not his portraits were all different. Warhol replied, "Why don't you ask my assistant Gerry Malanga some questions? He did a lot of my paintings."

Another well-known example is from Damien Hirst, one of the wealthiest and most famous living artists. Hirst is known for creating large volumes of art across a variety of mediums. Hirst's workshop apprentices painted his series of 1400 spot paintings, almost exclusively. In an article for The Independent, Hirst was quoted as saying, "The spots I painted are terrible; the best person who ever painted spots for me was Rachel. She's brilliant. The best spot painting you can have by me is one painted by Rachel."

Authorship and capitalism

One might wonder why an artist would subject themselves to receiving little or no credit for working on someone else's artwork. 

Lesser-known artists frequently accept apprenticeships with well-known artists to help advance their careers and learn the craft. For centuries, humans have contributed to artistic works without the expectation of authorship or recognition beyond a mere line on a resume. 

In modern democratic meritocracy, is the concept of getting paid a salary for one’s labor so abstract?

When we view art production as a brand, it aligns with structures already in place. Imagine a typical assembly line in a given industry, such as phone manufacturing. Hundreds of individuals are involved in various roles from the beginning to the end of the phone production process. Some may be responsible for designing the phone, while some assemble parts. Additionally, some individuals may negotiate funding, while others focus on sourcing materials, among other responsibilities. Many people play a part in the creation, but it is not logical to credit every worker involved for creating the final product. It simply doesn't make sense. And in many cases, workers sign away their legal right to the intellectual property they are creating.

How do we handle AI and authorship?

Society’s expectations regarding authorship can be quite absurd. For example, is it realistic for society to expect an artist to create, transport, and install large sculptural works, some of which may weigh several tons, entirely on their own?

Within the context of authorship, whether an idea is good or bad, we often feel compelled to share it with others. If our idea comes to fruition, even if someone else ends up doing the bulk of the work, we may still feel the need to take credit for the original idea. This desire for recognition is not necessarily incompatible with how artists, tech tycoons, and ghostwriters have historically been acknowledged for their contributions.

As a thought exercise, imagine a person visiting the Eiffel Tower and taking a photo of it. The resulting photo is almost identical to the thousands of photos taken by other tourists who visited the tower that same day. 

In this scenario, each individual who took a photo of the Eiffel Tower is credited as the author of their own photograph, even though their involvement may be minimal. The camera processes the photons on its sensor, and modern digital cameras use neural networks to instantly color correct the image, account for lens aberrations, and reduce noise. The camera does most of the actual work in producing the image of the Eiffel Tower, except for the act of clicking the button. 

In the same vein, a person can prompt an AI to create an image of the Eiffel Tower. The machine produces the image using neural networks and algorithms. The user has little to no involvement in the physical creation of the work, similar to how a photographer has little to no involvement beyond clicking a button. The user prompted the machine to create the work. The same way Damien Hirst promoted Rachel to paint his spot paintings. This concept of authorship is not outside the realm of how authorship is currently credited in the art world.

Camera information, such as the camera model and settings used to capture a photograph, are typically included in the "notes" or "description" section of a photo caption. This information can be helpful to viewers interested in the technical aspects of the photograph, or for photographers who may want to replicate similar settings or techniques.

For example, a photo caption may include the following information:

Title: Eiffel Tower

Author: John Smith

Medium: Digital photograph

Notes: Shot with a Canon EOS 5D Mark IV, using a 24-70mm lens at f/8, 1/100 sec shutter speed, and ISO 200.

Authorship: What’s the big deal?

If you think about it, your phone runs on computer programs and AI algorithms that handle basic tasks such as improving selfies. Your phone’s neural network helps it perform portrait mode, night sight, and other features. It automatically formats your photos and presents them to you in a format that you can share online.

Let's say you are using a popular social app and take a selfie. Before sharing the image, you decide to apply a puppy face filter. An AI recognizes your face and superimposes a cartoon puppy face on your selfie. You are still the author of your selfie, correct?

But maybe equally as important to consider, under the current framework, AI tools are owned by corporations. If we attribute AI as an author on a given artwork, what can of worms are we opening? Does it mean we need to pay royalties to corporations for everything a given AI creates? Do we think corporations would prefer that?

Photographers within the art world make a habit of challenging the laws. There are photographers who take photos right off Instagram and put those up in galleries. Some artists take pictures of the insides of their neighbors’ apartments and put those up in galleries too. 

To answer the question of AI and authorship, we must consider the long precedent of legal challenges and Supreme Court rulings that have been established for artists. In my view, this suggests that we should adopt similar principles to determine authorship in works created with AI assistance.

  1. For photos taken with a camera, the person who pressed the shutter button is generally considered the author of the photo, regardless of the level of involvement in the creation of the image by the camera itself.
  2. For works created by AI or machine learning algorithms, the person who prompted the AI to create the work may be considered the author.
  3. In photo captions, credit should be given to the machine in the "description," “technical info,” or "notes" section.

In addition, both Springer Nature and Taylor & Francis have issued statements requesting authors to clarify any AI interactions in the "methodology" or "acknowledgment" sections.

It's worth noting that the rules for authorship can vary depending on the context and the specific situation. These rules are general guidelines and may not apply to all cases.

Upgrade your document quality with Curie. Our AI-powered tool provides top-quality editing services for a variety of documents, ensuring clear and effective communication. Save time and avoid errors with Curie.

chatgptaiB2BAuthorshipEthics in Authorshipai and academics
Table of contents
FacebookTwitterLinkedInCopy linkEmail
Join the newsletter
Sign up for early access to AJE Scholar articles, discounts on AJE services, and more

See our "Privacy Policy"