Regenerating beauty in the age of AI

I wasn’t really looking for it.

I was just casually digging through folders on an old hard drive — the digital equivalent of a cluttered attic — hunting for something else entirely. But then there it was. A PDF with a filename I almost scrolled past: Beauty_thesis_FINAL.pdf. 🤔

I just paused for a moment before opening it. With a twinge of dread, I’ll admit.

There’s a certain kind of unease that comes with encountering your own thinking from 25 years ago. 😬 It’s not quite embarrassment, not quite pride. It’s closer to running into a former version of yourself at a party — someone who shares your face and your history but whose aspects you no longer fully inhabit, whose youthful idealism and earnestness make you a bit uncomfortable in ways you can’t entirely justify.

I opened it anyway. 🤠

The title page read: Design for Beauty: Exploring the Human Experience of Interaction Design. My Master’s thesis, submitted to the School of Design at Carnegie Mellon University, May 2001. Advisor: Richard Buchanan.

I read the abstract first, the way you might approach something cold — carefully, with a slight distance:

“Beauty becomes a vital issue as our environment of human experience becomes rapidly shaped by electronic, multifunctional products... beauty, as it pertains to interaction design, transcends mere surface ornament to become an emergent value of human-product interactions that operates in experience through human attention and emotional attraction.”

I had to stop there.

Not because it struck me as wrong. Because it struck me as true — and because I realized, sitting here with my laptop in 2026, that this is almost exactly the argument no one in design is making right now. At least, not in those terms. Not with that seriousness. Not as a foundational claim about what design is for. 🤨

The world that thesis was written into looks almost unrecognizable from today. The examples I used to illustrate certain aspects of beauty in digital interaction included experimental Flash websites, the Palm Pilot, designs from Sony, well before Apple became a consumer device kingpin. The World Wide Web (yes we spelled it out) was still young enough to feel like a genuine aesthetic frontier — a place where designers were seriously arguing about whether digital could be beautiful, not just functional, not just usable, not just “delightful” in the thin trendy sense of the word.

Those arguments felt urgent then; but somewhere along the way, they stopped being made.

I want to be careful here about what I’m not saying. I’m not saying design got worse, exactly, or that the designers working today are less talented or less thoughtful. The tools are incomparably better. The reach is vastly larger. The diversity of practitioners is broader and richer. Many things genuinely improved! 🌟

But something also got lost — or more precisely, something got substituted. The design field didn’t abandon its values so much as quietly replace them. In the years since 2001, design discourse — now amplified through LinkedIn “hot takes” and a dozen specialized publications that didn’t even exist back then — increasingly centered on craft. Pixel precision. Tool mastery. Figma fluency. Taste as a terminal value, something to be cultivated, displayed, and argued about in comment threads.

Craft isn’t wrong. But craft is how you make something.

Beauty — in the sense my thesis was trying to articulate — is why it matters that you made it at all.

When craft becomes the destination rather than the instrument, you get technically exquisite work that is spiritually empty. You get interfaces of breathtaking precision that leave users feeling vaguely depleted. Cold. You get “taste” as a competitive credential rather than a moral orientation toward the people you’re designing for. 😬

And “taste” — unlike beauty — has no ethical accountability built into it. You can have exquisite taste and still design something that diminishes the people who use it.

The examples are not hard to find. The Sonos app redesign of 2024 was visually coherent, and modern enough by craft standards — yet so broken in practice that it cost executives their jobs. Dark patterns saturating digital life are often beautifully considered at the component level, deployed precisely to work against the people using them. And AI-generated UIs now produce a confident visual adequacy — hitting every expected craft signal without the distinctiveness that comes from a designer wrestling with a real problem — that risks becoming a mediocre default aesthetic of a generation, because it offends no one and serves no one in particular.

Taste, in each of these cases, was present. The question of what the design was for — whose life it served, what human value it honored — was nowhere to be found.

This substitution of craft for purpose didn’t happen in isolation. The very concepts my thesis had treated as humanizing possibilities were being captured and turned against users at the same time.

Csikszentmihalyi’s flow — which I had used as a bridge between abstract aesthetic theory & tangible design practice — was being reverse-engineered by designers who understood its mechanics perfectly…yet deployed them against users, as manipulation, as gamification, as dopamine slot machines. The conditions of flow (clear goals, immediate feedback, the merging of action and awareness) were preserved wholesale. The life-enhancing purpose was stripped out entirely. What remained was compulsion dressed as engagement — the infinite scroll, the notification pulse, the algorithmic feed optimized not for enrichment…but for entrapment. 😑

This wasn’t an accident or an oversight. It was a deliberate design decision, made by people who knew exactly what they were doing, in service of business models that treated human attention as a resource to extract rather than a faculty to honor. The cynical capitalism that my thesis gestures toward only obliquely — the “grim consequences” of alienation, the “spiritually impoverished” society — turned out to be not a cautionary possibility but an actual destination.

We built it on purpose, with great craft, and called it product design. Ha! 🙃

And now we are in the age of AI, which has made all of this feel simultaneously more urgent…and more complicated. Generative systems can now produce interface designs, visual styles, and interactive experiences at mass scale. They can replicate the surface appearance of beauty — the aesthetic signals of care & refinement — while having no relationship whatsoever to the humanizing values that beauty, in any serious sense, actually requires. The question I was asking in 2001 — can digital be beautiful? — has assumed a darker shadow: can an algorithmically generated experience be beautiful at all? Or does the way it gets made foreclose something essential about what beauty requires?

I don’t have clean answers. I’m not sure anyone does yet, really…

But I do know we can’t begin to address those questions without recovering pertinent vocabulary that the field has allowed to atrophy — without taking seriously the philosophical traditions that tried to understand what beauty actually is, what experience actually is, what it means to design something that genuinely humanizes rather than merely optimizes or extracts. 🙌🏽

My thesis conclusion, written in May 2001, said this:

“To combat the ‘ugliness’ of dehumanizing, alienating artifacts that deprive humans of their ideational, cultural, and personal aspirations should be the driving motive for those who wish to create beauty in interaction design.”

Reading that sentence now, I feel the weight of how little it was heeded — and how much more urgent it has become. Because the 2001 version of me was framing ugliness as a moral category, not just an aesthetic one. Dehumanizing design wasn’t bad taste; it was a failure of ethical responsibility. That framing felt serious then.

It feels like an emergency now.

This 3-part miniseries is not an attempt to rehab a 25 year old document. That thesis had its limitations — it was written by a naive, idealistic grad student responding to a world that no longer exists, citing examples that now feel like time capsule specimens. But the questions it was asking didn’t expire with those cool Flash websites. They went underground. And this critical moment we’re in now — where AI can generate the appearance of thoughtful design faster than most designers can think about what thoughtful design actually means — may be exactly the moment to dig those questions back up.

As navigation, not as nostalgia. 🙏🏽

Next: Part 2 — The questions that got buried. The philosophical discourse that design stopped teaching…and why its absence is costing us now.