The Panic Before Acceptance: AI in Art and Architecture
I’m constantly talking with friends and colleagues in architecture, the broader design world, and business about artificial intelligence, and I am surprised by how emotional some of these conversations have become. A recent example hit close to home: my wife, who creates detailed digital artwork using a mix of AI tools and traditional software, saw a comment from someone who said she was furious that her mother had bought this “bullshit AI art” and called it ethically wrong. The commenter knew nothing about the process, the hours of iteration, or the years of visual training that go into the work. What she reacted to was the idea of AI itself.
That reaction is increasingly common, and it tells us something important: many people aren’t just critiquing individual pieces — they’re afraid that AI is cheapening creativity. As an architect by training, with executive experience in manufacturing, architecture, and construction and years of consulting for AEC and related industries, I’ve seen this pattern before. Every time a powerful new tool appears — photography, printmaking, CAD, BIM — there’s a wave of moral panic. The tool is framed as a shortcut, an enemy of “real craft,” or a threat to jobs. Then, slowly, serious practitioners figure out how to use it well, and it becomes part of the standard toolkit.
My argument in this article, aimed at architects, artists, and other design professionals, is straightforward: using AI in creative work is not cheating. It is the latest in a long tradition of technologies that extend our capabilities, change our workflows, and force us to rethink what “craft” looks like — without removing the need for a trained eye, professional judgment, and ethical responsibility. My goal is to separate genuine ethical concerns, such as deception and misuse, from an overall rejection of the tool itself, and to offer a way for design professionals to think about AI that is historically grounded, practical, and honest about both the opportunities and the risks.
New Tools, Familiar Panic
History is full of moments when new tools were treated as existential threats rather than as natural evolutions of practice.
When photography first emerged, many painters and critics refused to see it as “real art.” A machine, they argued, was simply capturing reality; where was the human skill? Over time, of course, we came to understand photography as its own medium, with unique aesthetics, constraints, and disciplines. Owning a camera never made anyone a great photographer. The craft moved from painstaking manual reproduction of scenes to composition, timing, light, and storytelling.
Later, artists like Andy Warhol and others in the Pop Art movement pushed authorship in another direction. They used screen printing, fabrication, and even outsourced production to studios and manufacturers. Critics bristled: if the artist’s hand wasn’t on every surface, was it still “authentic”? Today, we accept that conceptual authorship, direction, and curation are forms of creative labor. We judge the work as a whole, not whether the artist personally pulled each print.
Architecture has its own versions of this. In the Renaissance, architects began to shift from being primarily master builders — deeply involved in hands‑on construction — to designers and thinkers who coordinated teams to realize complex visions. That change was not universally welcomed. Some argued that if you weren’t the person laying each stone, then you couldn’t truly claim authorship of the building. However, 500 years later, the modern practice of architecture, engineering, and construction is built on specialization. In general, architects and engineers design and direct, while contractors and fabricators execute.
Fast‑forward a few centuries, and we see the same storyline with CAD and computer rendering. When CAD entered the profession, many feared it would “kill” the craft of hand drafting. But what it really did was move effort away from repetitive line work and toward coordination, detail, and iteration. Digital visualization prompted similar worries — would photorealistic renderings mislead clients or trivialize design? Over time, they became standard tools. Their presence didn’t eliminate bad design, nor did it prevent good design. They changed where and how the work happened.
AI now sits squarely in this lineage. The pattern is familiar:
- A new tool appears.
- Early use cases are clumsy, gimmicky, or dishonest.
- People extrapolate from those examples and declare the tool itself “fake,” “cheating,” or “unethical.”
- Serious practitioners quietly learn to use it well, and the tool becomes part of normal practice.
AI as a Tool, Not a Shortcut
One of the biggest misconceptions about AI is that it “does the work for you.” In reality, most raw AI output looks like exactly what it is: a first draft.
One of my favorite Ernest Hemingway quotes is, “The first draft of anything is shit.” I believe that applies to writing, architectural design, sketches, and AI. The difference between a forgettable AI image and a compelling piece of art, or between a generic AI massing and a viable building, is what happens after that first draft.
Most design practices already work with layered authorship. In an architecture firm, for example, junior designers or interns often generate early concepts. More experienced designers and technical staff then critique, redirect, and refine those ideas with a better eye for composition, context, structure, and constructability. The final design is rarely the untouched product of a single hand; it’s the result of iteration, feedback, and synthesis.
AI can slot into that process much like a very fast, very capable intern. You give it direction, constraints, and references; it provides options. Some will be useless, some intriguing, and a few worth pushing further. But nothing it generates is “yours” until you’ve reviewed, edited, and integrated it — and taken responsibility for it.
That responsibility is especially important in fields like architecture and engineering. If AI helps generate a detail, diagram, or narrative, it still has to pass through human review. Once you sign and stamp a drawing set or present a concept to a client, you own it. Blaming the tool will not save you if there are errors or omissions.
The criticism that AI “makes it too easy” mirrors complaints about every productivity‑enhancing tool. Cameras made capturing images easier than painting; plotters made producing multiple sets easier than hand‑inking; BIM made coordination easier than manually chasing down every clash. But ease of execution does not negate the need for vision, taste, and rigor. It simply changes where the hard work sits. Instead of spending days making 50 physical study models, we can now generate dozens of digital variations in minutes and devote more energy to selecting, combining, and refining.
“Prompt‑dump” work — the kind where someone types one vague prompt, posts the first output, and calls it finished — usually looks like exactly that: a first draft pretending to be a final. Professionals can see the difference. The real value, as always, lies in the quality of judgment applied after the tool has done its part.
Lessons for Architects and Design Professionals
In the AEC world, skepticism about new tools is practically a rite of passage. Older generations have seen drafting boards replaced by CAD, CAD by BIM, and hand‑built chipboard models by digital massing and VR. Each shift has sparked fears of job losses, lost craft, and lower standards. There is some truth underneath those fears. New tools do change staffing models and required skills. In the 1960s, assembling a drawing set might require teams of draftsmen. Today, a much smaller team can produce a more complex set of coordinated documents. That has real implications for how many people a firm needs and what their roles look like.
But these shifts have also created new possibilities. As documentation has become more efficient, architects can spend more time on:
- Design exploration and iteration.
- Integration of sustainability and performance considerations.
- Coordination with increasingly complex systems and stakeholders.
- Communicating ideas to clients and communities.
AI has the potential to extend that trend. Thoughtfully used, it can:
- Generate early concept options for unpaid competitions and proposals that have historically drained time and resources.
- Suggest variations of forms, facades, and layouts that might not emerge from habit alone.
- Draft first‑pass details or specifications in areas where firms already copy and paste from past projects.
The key is not to skip the review. I often compare AI to an excellent intern: it can get you to a strong starting point quickly, but you still need to check everything. We’ve already seen in other professions what happens when people forget that. In law, for example, there have been cases of attorneys submitting briefs with AI‑invented citations. They didn’t look foolish because AI exists; they looked foolish because they failed to verify the work they put their name on.
Get Brad Riemann’s stories in your inbox
Join Medium for free to get updates from this writer.
For architects, the stakes include not just reputation but liability and public safety. Using AI to assist with drafting or detail generation can be powerful, but only if every output is treated as a suggestion, not a finished answer.
What’s actually unethical — and what isn’t
It’s important to separate two very different conversations:
- One about how AI is built and deployed (training data, consent, compensation, environmental impact, misinformation), and
- Another is whether using AI as a tool in your creative process is inherently “cheating.”
The first conversation matters a lot. There are valid concerns about models trained on artists’ work without consent or compensation, about using AI to imitate specific artists too closely, about deepfakes and political misinformation, and about the energy consumption associated with large models. These are real issues that deserve serious attention, regulation, and professional standards.
But they’re not the same as saying, “If you use AI at all, your work is ethically wrong.”
From an ethical standpoint, a few principles go a long way:
- Honesty in representation: If you market something as “handmade” or “hand‑crafted,” it should be exactly that. If your process is digital and tool‑assisted, you don’t need to list every tool, but you shouldn’t deliberately mislead buyers.
- Responsibility for what you put into the world: Using AI to fabricate images or videos that depict real people doing things that never happened is a very different category than using AI to help generate a building massing or a piece of abstract art.
- Respect for others’ work: Using AI to explore ideas is different from using it to closely mimic a specific artist’s style and selling the result as your own invention.
Most architects, designers, and artists I know are not trying to scam anyone. They’re trying to navigate a changing landscape, using new tools to do better work, more efficiently, for clients and audiences who expect high quality.
A more constructive way to think about AI
When someone says, “Using AI for art is cheating,” I’m tempted to respond with a simple analogy: Is using photography to capture an image cheating? I can sketch reasonably well, but my sketches will never be as accurate as a photograph. Yet we don’t accuse photographers of cheating because they didn’t paint every pixel by hand. We recognize that photography is its own medium, with its own skills and standards.
AI is on a similar trajectory. It:
- Makes it easier to explore a wider range of options quickly.
- Allows us to offload some repetitive or low‑value tasks.
- Demands that we develop new skills in prompting, curation, and critical review.
It does not:
- Remove the need for a trained eye.
- Eliminate professional judgment or liability.
- Automatically turn bad ideas into good ones.
My hope for architects, artists, and other design professionals is not that they embrace AI blindly, but that they recognize their skepticism as part of a long, familiar pattern. Every major tool that reshaped our disciplines went through a period of backlash and doubt. Every time, the real question turned out not to be “Is this legitimate?” but “What does good work look like when this tool exists — and how do we keep our ethics intact while using it?”
AI doesn’t erase the need for craft, discernment, or responsibility. If anything, it makes them more visible. The gap between thoughtful and careless use is obvious.
So instead of asking, “Is AI cheating?” a better question for our professions might be:
How can we use AI in ways that strengthen our work, respect our clients and audiences, and align with the values we want our practice — and our industry — to stand for?
