Profitable Aesthetic — Domination of the Elite,
by the Elite, for the Elite

My impression of these two images was that they looked like something you’d see on booking.com or Airbnb. Both images are highly realistic, capturing the detail and representing the typical imagery of Norwegian and Japanese traditions and nature. They visualise the romantic atmosphere likely used in tourism. Yet, these places don’t actually exist. So what exactly is the purpose of these images?

These images might be used as demo shots for some kind of proposal or mood board. Or perhaps a travel agency could use them as image advertising for a destination. I don’t really understand what purpose such photos serve, or what effect they are meant to have. But it’s clear who will evaluate them. They are intended for the wealthy elite living in the Global North. To my eyes, these images appear to be nothing more than a new profit-making tool designed to capitalise their romanticism and capture their attention.

Design and aesthetics reflect the social phenomena of the world we inhabit. And there is a history in which the aesthetics favored by elites living imperial lifestyles have remained dominant. The modern AI industry faithfully follows this manner, neatly extracting aesthetics likely to appeal to Global North markets. While notorious for collecting vast amounts of data without permission, they meticulously analyze the romanticism of the Global North, further rank these data of aesthetic, and brilliantly meet those expectations.

Fast food, Fast fashion,
Fast graphic

Generative AI has become incredibly capable at generating realistic images. It processes images at astonishing speeds, producing professional-looking photos and logo designs. This tool is truly trendy. That sense of convenience at its speed feels strangely familiar. Yes, fast food and fast fashion.

I see generative AI as something like fast food or fast fashion for graphics. It’s incredibly convenient, and if I had to describe the experience, I’d say it’s “so cool.” Its speed and affordability are appealing. With Midjourney, paying 120 kr a month gets you stylish, trend- reflecting designs instantly. It even makes me wonder if, rather than studying design techniques, it’s better to cultivate that eye for foresight and evaluation, train your eye to discover new things, and learn to use generative AI creatively as a collaborator. However, looking closely at what’s happening behind the scenes (and what’s likely to happen next), the situation resembles the fast food and fast fashion industries in many ways. Key points include “environmental destruction,” “labor exploitation,” and ultimately, “harm to well-being.”

Environmental Destruction

The development of AI and global demand are driving the need for more massive data centers around the world. OpenAI, Nscale, and Aker have begun construction on Europe’s largest artificial intelligence data center in Narvik, equipped with 100,000 GPUs.¹ Northern Norway is experiencing a surge in requests from tech companies seeking to build data centers, drawn by its cheap and clean electricity.² OpenAI describes the massive Narvik data center project, known as “Stargate Norway,” as playing a crucial infrastructural role in meeting Europe’s AI demand.³ The energy consumption of this data center is enormous. Long-term plans call for 230 MW of usage, with potential expansion to 290 MW. This equates to the annual electricity supply for over 110,000 households.⁴ With Narvik’s population at approximately 21,000, this means a single data center will consume over 5.5 times (even more) the electricity used by the entire city. You might think that since northern Norway has surplus electricity, it’s not such a big deal if electricity prices rise fivefold. But that is precisely the problem. That’s how AI companies pursue infinite expansion in nature and locals.

Norway is becoming an AI Station to supply cheap electricity, but what existed before the data centers were years of preserved nature and animal habitats. If AI development continues at this pace, there will be no end to data center construction. So what exactly do the GPUs in these data centers primarily process? They likely will handle mostly generated fake images and, generated pornography, and trivial social media content and how much of this is actually intended for public benefit is questionable.

Labor Exploitation

Labor exploitation is indispensable in generative AI development. This includes not only the exploitation of designers and illustrators, but also the exploitation of content moderators working in the Global South. AI companies justify training their AI by using works found on online portfolios and social media without permission, claiming they are in the public domain.⁵

But behind this, something is being sacrificed. Why is it that even if you input “Genocide” into generative AI, the system won’t produce realistic images of corpses? It’s because real human beings taught the AI not to do so. The job of teaching AI “What is a corpse?”, “What is rape?”, “What is cannibalism?” is called content moderation. Many employees working as a content moderator at an outsourcing company in Kenya suffered trauma after reviewing thousands of images depicting suicide, murder, and rape, one by one. This included images of child rape and child murder. Many who took this job still struggle with PTSD, suffer mental breakdowns, and face consequences in their work, relationships, and family life. The hourly wage for this work was less than $2.⁶

The development of generative AI is built upon their sacrifices. How often are these stories even told? Hearing this story, I couldn’t help but recall the Bangladesh factory collapse where fast fashion companies had outsourced production.

Harm to Well-being

Unfortunately, the impact on well-being is often only recognized long after the trend has emerged. This seems to be what fast food and fast fashion have shown us. It has been three years since ChatGPT launched, but recent research conducted by MIT suggests ChatGPT has an effect that weakens critical thinking.⁷ In this study, students were divided into three groups for an essay-writing task: one group wrote using ChatGPT, one group wrote using a search engine, and one group wrote using only their own brains without any tools. What these three groups revealed was about brain activity. The group using ChatGPT showed the least neural connectivity response in the brain compared to the other two groups, indicating the lowest engagement. Furthermore, their essays contained highly homogeneous sentences.

While this study does not present simplistic findings that using ChatGPT makes you stupid, considering the reduced brain engagement it demonstrates, we cannot definitively rule out the possibility that AI may have adverse effects on our cognitive functions in the future.

When fast food first appeared, its extent and nature as a health hazard only became clear later. There is “time lag.” When fast fashion emerged, the psychological and social harm it causes by fueling relentless consumption and material desire, along with the fact that microplastics like polyester pollute oceans and bodies, also only became apparent later. Yet somehow, these issues were overlooked. Why? Because the research findings aren’t immediately apparent when these industries first appear. We are essentially being subjected to a massive social experiment by these industries. The harmful effects only start to be pointed out years later, and by the time they become clear, both individuals and society are already completely dependent. We are like drug addicts who know better but still can’t resist. Understanding the harm Fast Graphics inflicts on our well-being will likely come a little further down the road. By then, we will surely be utterly dependent.

AI and Capitalism are Robbing Designers and Illustrators of Their Democratic
Time and Rights

We seem to have witnessed capitalism stealing time from democracy. The AI industry is the prime example. Despite numerous global issues demanding our attention, we are too busy to address or discuss them. Isn’t this a paradox? Indeed, we are forced into lifestyles that prioritize time spent on capitalism over time devoted to democracy. With the emergence of generative AI, it seems that designers and illustrators engaging in democracy are getting into an insecure position.

As generative AI technology advances, designers and illustrators will likely see their freedom and rights steadily eroded. What if you insist that your work not be used to train AI, or claim that generative AI constitutes plagiarism, or argue that using generated images and videos feels uncomfortable from an environmental perspective? While such claims might seem acceptable now, we don’t know how the future industry will respond. People around you might dismiss your personal views without respect, labelling you “outdated,” “stubborn,” or “hypocritical.” They might push you aside with “If you take on that stance, it’s your own fault or responsibility,” leaving you isolated. In this way, a democracy whose time has been stolen by capitalism lacks the capacity to protect what you hold dear. This is how time and rights are taken away.

Saying “No” is Not a Shame

By the way, we could all become climate refugees. Or we might have to be the ones accepting climate refugees. Despite this, Sam Altman is saying leisurely things like AGI or new technologies will solve the climate change problem. But, climate change isn’t necessarily a problem that can be solved by technology alone. We already know many solutions: shifting to plant-based diets, using public transportation, banning private jets, limiting air travel, sharing appliances etc. These are rather political and social problems than technical problems.⁸ It’s simply a matter of us choosing them. And frankly, this AI hype has created tremendous waste for us who must address the urgent climate crisis. Nations must stop fantasizing about technology. As long as technology is developed under capitalism, it will only become more violent. What we need now is not technological fantasy, but the courage to speak up and system change. Opposing generative AI or saying NO to expectations for new technologies is neither naive nor shameful, especially in this society where capitalism is running out of control.

With Self-recrimination — Hypocrisy and Dilemma

Since I began researching AI as my PhD subject, I’ve found that the problems with generative AI are pretty similar to those of fast food and fast fashion. This realization made me want to share my thoughts through an outlet like this, YMT. But that doesn’t mean I’m perfect, of course. Some might point out that this essay itself is likely written using information found scattered across the internet. Some might even question whether I needed to use Grok to generate examples in the first place. People might call me a hypocrite. And I must confront that dilemma and hypocritical part of myself, because I stand in a privileged position. I have opportunities to publish, present at conferences, exhibit work, and share my ideas where people will listen. It’s like having a slightly larger megaphone than others. That is precisely why other’s criticism that my words and actions remain hypocritical does not justify stopping the pursuit of knowledge.

My position comes with the responsibility to keep learning and to share what I learn for a more fair society. I hope you too will use the megaphone of your own position to raise your voice.

(This essay is not written by Generative AI.)