The aim of this piece is to inform the reader of the impact generative AI has on artists and other people in the creative industries, through job security, environmental impact, and copyright concerns.
What does it mean to be human? Most people would agree that it isn’t just eating and sleeping. In my opinion, our desire to create is an essential part of what defines us, and it has certainly led to our prevalence on this planet. Why then does it seem we are so set on stripping away this very fundamental trait? AI-generated art is becoming a gargantuan problem in today’s society. It destroys artists’ jobs, breaches copyright laws, and dumps greenhouse gases into our atmosphere. Overall, AI art generation is a poison, to be handled with the utmost care.
Firstly, generative AI threatens the job security of artists. The financial side of art is diverse and varied. Income generation can range from commissions by individuals to corporate contracts, but all are facing a rapid decline. With the advent of image generation, many major movie studios are incorporating more and more AI-generated elements into both their films and promotional materials. The effects of this are already evident, with thousands of artists left without work or a method of making money. Independent artists have also reported a massive decrease in commission rates, almost certainly influenced by the rise of easily accessible free image generation. After all, who would pay money for something you can just conjure up in 20 seconds? So why do we need to keep artists around? Simple. Beyond the implicit harm of destroying jobs, the art itself is more than just a picture; it provides a message—a meaning. With AI art, we all but strip this out. There is no thought behind every meticulous brushstroke, only the watered-down ideas of others. It rips identity from art and inhibits the use of art for progress, protest, or deeper thought.
Alongside this, the use of these AI tools requires vast swathes of resources. According to OpenAI’s publicly released statistics, ChatGPT is responsible for over a thousand tons of carbon dioxide released every single day, roughly equivalent to 600 airplane flights. Not great. Alongside this, text generation is one of the simplest, cheapest forms of neural network. Image and video generation are tens to thousands of times more hungry for power. A single video generated by Sora emits many times more CO₂ than a ChatGPT prompt. When you consider the volume of use these products see, it can become a major source of environmental harm. In fact, data centres account for nearly 3% of global carbon emissions. AI is not the only thing running on these servers, but it is a very substantial part of it. It’s safe to assume that AI is responsible for over 0.25% of net emissions, which is an absolutely horrific amount.
There is also the issue of ethical training. ChatGPT is trained on unfathomable quantities of text, all mercilessly ripped without consent from the internet. If you’ve ever contributed anything online—a Reddit post, a comment on YouTube, just a couple of messages on MySpace—chances are that inside ChatGPT is a little bit of you. This issue is already troubling when it comes to text generation: anything produced is just a rehash of existing information. Where it gets cripplingly bad is with art and music. AI-generated art is notorious for bearing a very…striking resemblance to actual works. One infamous case of this is with Greg Rutkowski, who said,
“It’s been just a month. What about in a year? I probably won’t be able to find my work out there because [the internet] will be flooded with AI art.”
This was in mid-2022, and over the past three years things have not improved in the slightest. In fact, the scope of his statements has extended dramatically. Now musicians, photographers—really any creative industry—is at threat of non-consensual data harvesting. Even when it isn’t immediately obvious, everything generated is just stitched-together chunks of other people’s work. Nothing is original.
Overall, generative AI has its uses. Enhancing aerodynamic efficiency in planes, protein folding, and many other fields can leverage AI to rapidly accelerate progress. However, art is a field where the application of AI has been decidedly detrimental. It has stripped works of originality and soul, producing endless rehashes of stolen artwork, all while releasing massive amounts of greenhouse gases into the atmosphere. Our history has been shaped by art for millennia, but with the advent of AI, perhaps our future won’t.
Open Event Registration - Interested in enrolling?
Employment opportunities - click if you're interested in working at McGlashan.