“Hey, I Drew That!”: AI’s Impact on Artists

Have you ever been creating a graphic and wished you had an image, but it doesn’t exist? Are you too busy to take your own photos? Then you were born in the right generation, because generative artificial intelligence is advancing at speeds like never before. Programs like DALL-E, ChatGPT, and Midjourney have released features that allow the user to type in a prompt and receive an image matching exactly what they asked for.

Well, hopefully.

The internet has been flooded with AI generated images of people with three arms, dogs melting into each other, stairs to nowhere, and an abundance of other blunders. While these mistakes may have been easy to spot at first and immediately indicated to the viewer that the image was created by AI, as the systems update and become smarter it is becoming more difficult to differentiate between human and machine made work.

💭 Where does AI art come from?

According to Adobe Firefly, “AI art is, simply, artwork made with the assistance of generative AI — a technology that finds patterns in big datasets and uses that information to create new content.” This means that each image that AI generates is composed of thousands or even millions of preexisting images, photographs, and text that can be found online. This creates an issue for artists who have their work published online, because their art is being used without their knowledge or permission. There is also not currently a way for artists to be paid for their work that is contributing because of the size of the datasets that the AI pulls from. There is no system in place for artists to receive royalties or credit for their work because it is simply not being tracked. This is especially harmful for smaller, independent artists. They are missing out on the potential income that could come from large companies using their work. Despite some websites allowing users to specifically opt out of AI scanning, such as DeviantArt, these still usually include a “no guarantees” policy. Some of these websites only implemented these policies after huge public outcry from their user bases, especially those that catered to artists. This doesn’t even begin to cover websites that encourage the use of AI or even have their own built into the platform, such as Instagram and Facebook. 

🎨 How is it affecting artists?

AI generated art becomes a huge issue for artists’ income streams because of its low cost. Many programs are cheap or even free. This means that less and less people are commissioning artists to draw their characters, for photographers to capture moments, and graphic designers to create visuals. There are even TikTok filters using AI that let you reimagine yourself as a Disney character, the opposite gender, and endless other expressions. Filters like these might replace going to visit the caricature station at a local fair, or doing a themed photoshoot at a local business. 

Being able to easily create images with just a prompt can also stifle creativity. When someone might have rolled up their sleeves and learned to use Canva or a similar application to make a flyer for their club, busy students are increasingly using AI for a quick and easy way to get the word out. Even companies with the resources to hire artists have been using AI for social media content creation, including the University of Delaware. While it may seem like a simple solution in the moment, in the long run it hurts artists who are not being utilized or compensated.

It also calls copyright laws into question, considering there is often not permission being given for use. However, because it is not the exact art being reused, there are usually no clear violations being made that would hold up in court. This means artists cannot legally advocate for themselves. There is a possibility down the road that new legislation could be created to address the innovation that has been made in the AI and technology space.

🧐 Why does it look like that?

AI art is notorious for not always creating what the user might have hoped for. From miscommunications to extra limbs, these mistakes usually make it abundantly clear that an image was created from a mix of other images. 

One of these miscommunications, as seen above, can occur when there are multiple meanings of words or phrases. When the user prompted the program to create an image of a fisherman, as in the occupation, it came back with a fish-man hybrid creature. This means the user needs to revise their prompt. These are bound to happen due to AI programs’ general inability to understand nuance and tone. Despite the artificial intelligence, it might not have the emotional intelligence to use context clues and understand when someone wants to see a person with a fishing pole rather than a sea monster. This can make the time-saving aspect of generative AI almost redundant, since the prompts often need to be revised in order to create a better result.

👥 Who else has AI art impacted?

Another side effect of lacking emotional intelligence is that AI art has to borrow it from the sources it pulls from. Similarly to text-based AI, if the information is biased a certain way, then the image it produces will have the same bias. This means that content containing misinformation, hate speech, or other harmful information can be used to educated AI models. AI art has been shown to disproportionately distort Black people’s facial features and hair textures. It also often creates people who are stereotypes of one or multiple of their identities, for example creating images of Black people who are criminals, or gay women who look masculine. This reinforces already existing stereotypes and perpetuates the mistreatment of many marginalized groups. 

🌎 How will AI art affect us in the future?

Generative AI uses an immense amount of computational power to train its technology and generate information. According to Noman Bashir, a Computing and Climate Impact Fellow at MIT Climate and Sustainability Consortium, “The demand for new data centers cannot be met in a sustainable way. The pace at which companies are building new data centers means the bulk of the electricity to power them must come from fossil fuel-based power plants.” This means that training new AI models rapidly becomes unsustainable, and a huge source of pressure for power grids. In 2021, scientists from Google and the University of California at Berkeley estimated that the training process for AI alone uses the same amount of power as 120 average United States homes in one year. This becomes more concerning when you look at how much the use of AI has grown since 2021, as well as the amount of programs that have been developed since then. 

AI art has also sparked conversations about the ethics of computer generated art and its effect on creativity. Some believe that AI art will work alongside art made by human beings and supplement it rather than replace it. The same way that modern art was once met with heavy skepticism, with many questioning its legitimacy, some theorize that AI art will have a similar journey in the public eye. The definition of art differs greatly among professionals and applies to this issue as well.

Conversations about AI will continue to evolve, with strong opinions on both sides. Some artists believe that using AI to create images could be ethical in a world where artists are highly valued and well-paid, but doubt that our society could reach that point. Others are completely against any use of AI, including using tools in graphic design programs like image expansion. The topic has stirred up controversy on social media, with commentary from artists and non-artists alike. Generative AI is a powerful tool that could quickly become unwieldy if not used with caution.

Written by June Reyes