From a designer’s perspective, AI is like the new version of plastic - distorting everything, at great ecological cost
Another stirring cross-post from Dieline magazine, from John Kazior, “a writer covering ecology and design in the climate crisis”, art director at the literary magazine the Drift and co-founder of Feral Malmö, a studio for urban-ecological practice and culture in southern Sweden.
This is a trenchant piece of AI-sceptic critique, from the perspective of the human designer:
How AI Cheapens Design (At a Great Ecological Cost)
It doesn’t come as much of a surprise to learn that, environmentally speaking, AI is an extremely wasteful and destructive technology. It’s rare that you can nail down the ecological cost of an image.
Still, researchers have assessed that generating a single AI image requires the same amount of energy needed to fully charge your iPhone battery. Training an AI model also requires “626,000 pounds of carbon dioxide, or the equivalent of around 300 round-trip flights between New York and San Francisco—nearly five times the lifetime emissions of the average car.”
This technology, marketed as artificial intelligence, also consumes millions of gallons of clean water, which would otherwise be utilized as regular old drinking water. One report found that, in a few years, it will demand the equivalent of half of the water the UK uses annually.
AI’s demand for perpetual access to consistent, maximum energy makes using renewables like wind or solar unpopular among tech companies, as the power supply depends on weather factors.
So, nearly all of the data centers fueling AI operate on fossil fuels—methane and coal. On top of that, the breakthrough technology uses and dispenses harmful chemicals like lead, mercury, and cadmium (which have been tied to soil, air, and water contamination).
Some consider this sort of ecocidal engineering to be “scary but exciting,” and we should be “grateful” for such technologies. I expect that for most people who live on Earth, these observed and documented facts are pretty much enough to regard proprietary AI technologies as a net negative for you, me, society, and certainly for the planet.
But for the sticklers out there who think Adobe’s Generative Fill tool is pretty neat or still buy the good ol’ “ends justify the means” argument—that the real and immediate evils of this technology are worth it to arrive at the purely theoretical goods of the technology—there are even more reasons to be against the adoption of AI into the design industry.
Because, like water, land, rare earth elements, and energy, AI needs to run, design, images, and the creative content made by designers every day are also among the resources consumed to keep these computational machines running. And the ecology of the AI industry can show us just how bad proprietary AI technology is for the commercial art and design industry.
After all, why is it that companies like Google, OpenAI, and Midjourney can siphon off so much fresh water and other resources for AI, seemingly without much thought of the costs, as parts of the United States and large parts of the world face water shortages?
You would think that because clean water is scarce and urgently needed by humans, it would be prohibitively expensive to use gallons upon gallons of fresh water to run a machine that makes some of the worst art you’ve ever seen. But in truth, to wealthy companies with access to vast land areas and geographies, all of it comes cheap.
“Cheap” not in the sense that it’s a good deal; cheap in the sense that it has been made easy to use and consume without much accountability or oversight. Raj Patel and Jason W. Moore’s book The History of the World in Seven Cheap Things explains the idea of “cheap nature” with the example of chicken, the most common bird and popular meat on the planet.
“Today’s birds are the result of intensive post–World War II efforts drawing on genetic material sourced freely from Asian jungles, which humans decided to recombine to produce the most profitable fowl. That bird can barely walk, reaches maturity in weeks, has an oversized breast, and is reared and slaughtered in geologically significant quantities (more than sixty billion birds a year).”
Further, Patel and Moore write that “the most iconic symbol of the modern era isn’t the automobile or the smartphone but the Chicken McNugget.”
Some have already begun to say AI is the digital version of plastic. But we may also think of AI design as tech’s answer to the chicken. The data, the visual creative information fed into these machines, is largely free to these companies.
They’re so self-assured in stealing creative content that they don’t even pay attention to the machines as they do it (to be sure, AI offers many exciting new opportunities regarding pleading ignorance).
An incalculable amount of time and work put into creating digital logos, graphics, animations, illustrations, setting type, taking and adjusting photos, producing icons, and styling layouts is consumed for free as the raw material for machine-produced design.
OpenAI quickly realized how cheap design has become through machine learning, as the first company they acquired was a design company. Now, AI has already infected everything in the design world, from interior, fashion, and architecture to UX, industrial, graphic, and packaging design.
There was little chance of resistance when programs like Adobe, Canva, and Figma were more than happy to jump on board by integrating their own tools.
But this isn’t fast food we’re talking about, right? Surely a robot can’t create a work of art! In terms of producing art, generative AI is overwhelmingly pretty bad and definitely not as good as the real thing. I could argue it’s objectively bad because AI is computational and unconscious.
Despite its marketed name, it’s very much still a computing machine. Its subjects, compositions, and style are not imagined, dreamt, experienced, believed, or hardly even created—they’re an averaging of data input. This process reduces the artwork of others to primitive, mechanized form factors dictated by the medium (browser, social media, phone, or computer screen) in a process where not producing something is unacceptable.
The only actual good art in the equation of AI was created by others, i.e., the raw data it trained one, and the output of the machine offers an image that attempts to cobble together an image or video file in response to a human’s prompt.
The output isn’t so much a piece of artwork but GUI, making AI a useful computer system made by and for people who mistakenly believe art is something that can and should be quantitatively defined through universal, observable signifiers.
In reality, “good” art—the best art—is defined by how it cannot be measured or distilled. In short, AI art is the unconscious reduction of good artistic works, processed into a file format to become something that’s nearly the antithesis of art, an assortment of data that’s supposed to look good.
Nonetheless, some people are dead set in their belief that AI is capable of making good art. Culturally, art is in some way always protected by the fact that the viewer defines art—and its value—subjectively. The protection subjectivity offers can make it difficult to cheapen art (not that there aren’t those who are trying.)
You can’t say the same about design and other commercial arts. Annoying as it is, the value of design is in significant part linked to commerce, of course. A good graphic designer has an eye for what makes type look harmonious and readable, and they create packaging that sells units for a company.
A good product designer understands ergonomics and ease of use, functionality paired with aesthetics, so an appliance will sell in an appliance store. We all understand how this works. Despite this glaringly obvious fact, the work of designers and the jobs of designers are under direct threat from AI companies.
The auto-branding nature of AI, which allows it to ceaselessly construct its own identity through content (content that often verges on the horrifying, but content nonetheless), primes it to undermine the design industry and its merits.
It’s creating content that’s “good enough” for corporate management, executives, “the algorithm,” and, of course, a global majority of consumers looking more at the price of that milk carton than they do at the design of the label. That is already proving itself to be true by the flood of AI-generated material on Pinterest, Instagram, TikTok, and all the other social media platforms.
Creatives are engaging with it and seeing it themselves. According to a recent survey from Its Nice That, 83% of designers already use machine learning tools.
Some notable details in the survey reveal just who is driving the usage and adoption of AI tools in design. Primarily, digital designers and strategists who work in-house and are the head of their team or department are the most interested and active users of AI.
They also tend to be older. The younger designers, who work as freelancers—think illustrators and motion designers (who are also generally pretty ticked off about big AI using their work as training data without permission)—are the most resistant and concerned about AI infringing on design. And a bit more than half of designers (58% overall) believe there will be no escaping AI design in five years.
Still, many designers say AI will just be another tool in the toolbelt: the designer will adapt, and, hell, it could even allow designers to focus more on the “real” creative work while leaving mundane tasks to the machine.
But these arguments could only be made by a designer who practices in a genuinely lofty position, somewhere high up in the hierarchy of an in-house brand, or maybe a star designer at a signature agency with deep-pocketed corporate clients (or they might just be designers who are simply too online and desperately need to go outside).
After all, who could look at the real, burning world we inhabit and believe we are on a gilded path toward a technological utopia? By this android-dream logic, the computer and the internet saved the publishing industry, and story illustrators are flooded with work, living their best, fulfilled creative lives (and probably making a nice living off of it, too).
By this logic, typography must be a common career path—there must be foundries in every town (or the online equivalent) making all sorts of rich new cultural movements through font and letterform.
Packaging designers must have so much power in the material and forms their design takes; they’re no longer shackled by the mundane tasks that limited their creativity before the emergence of the Adobe suite. In fact, every designer today must be living a life free of stress and is entirely creatively fulfilled.
Such arguments, and the myopia they imply, are insulting. But they are right on just one account—AI is a tool. It’s a tool that takes away the power of actual people to participate in creative work. It’s a tool that will rob illustrators of their work and prevent them from being able to feed themselves and their families from their work.
It’s a tool that can let brand executives assign what they see as tedious intricacies of typography to a machine rather than a passionate person who reserves their creative mind and hands for letterforms.
The nameless “designer“ that AI’s defenders talk about may indeed adapt to an industry happy to burn as many fossil fuels and pollute as much water as possible to churn out AI content. It’s just people who will lose their jobs.
Until one day in this Darwinian designer’s dream, there will be only one—the last true designer working the content production line, staring at a screen all day and night, nodding their creative little head to approve that each new piece of AI junk that spits out of the machine is indeed design.
Many environmentalists in the US thought that computers and the internet would lead to some new environmental salvation. Remarkably, there are people today who also believe that AI—seriously, this time—is going to be the tool that saves the planet and not, once again, spark an acceleration of consumption and extraction of nature and people all over the planet and fuel acute resource shortages.
(It’s almost as if these miraculous tools exist in the kind of society where powerful people would prefer to use and develop any and all tools they get their hands on for the express purpose of building up their fortunes and not for the pursuit of creative or ecological restoration).
Sadly, this new electric-powered illusion is buckling under the weight of its own pretense. AI is already feeding an imminent energy crisis for the sector. Ecologically and historically, this tends to result from making things like energy cheap—we are living through the proof of it as a global climate and ecological crisis.
The capabilities of AI and what it can do for design and the climate are undoubtedly overhyped and oversold, as pretty much all tech products are in our current culture. Their extreme resource demands make the suggestion of a fully automated society–or even design industry, a fever dream.
Unfortunately, that doesn’t mean the owners and developers of these technologies won’t keep making it worse for everyone who uses creativity to make a living on planet Earth.
And after all, if Midjourney and OpenAI’s investors are willing to cheapen and exploit clean air, water, rare earth minerals, and of course, a stable global climate to make shitty Lego web games, do you honestly believe they’d really stop for a graphic designer?
More from Dieline magazine - subscribe here