It is the very familiar things in life which are often the most difficult to define ... So it is with the concept of craftsmanship. We know what we mean by it, or seem to do so. We can call to mind examples of it and we can recognize it when we see it, or so we think. We speak of good or bad craftsmanship as if we had built-in standards and criteria ready to hand. And the closer we look at it the more complicated it all becomes.
–Harold Osborne, The Aesthetic Concept of Craftsmanship
As more research, explorations, and speculations into “the future of UX” (and AI’s role in it) come online, one idea is continually held up, like a shield, in front of the practice of designing the software interface. That’s the idea of “craft” as a central pillar of the discipline—a singular differentiating factor between human- and machine-driven digital production. And the more I see craft mentioned as a power of the designer, the more I feel that a definition or working understanding is urgently needed. After all, as I wrote in response to Figma’s “make designs” feature back in late June 2024, if we can’t define a term, we probably shouldn’t be resting our careers on it.
At the time, I also highlighted a prevailing feeling that design had, at that exact moment, been commodified; a notion that had, at best, arrived too late. The act and process of design (and the entire system of design as a phenomenon) has been commodified at least since it became uncoupled from the direct production of objects. With the introduction of mechanized mass production, the disciplines, actions, and practices required to produce material culture were gradually separated and specialized. Industrial design, machine design, engineering, and operation became discrete functions and those occupying each function could be (and were, and still are) swapped in and out at the will of whatever organization had hired them.
Harold Osborne, in an essay for the British Journal of Aesthetics (a publication I was excited to learn exists) is far from alone in identifying this moment as key to the changing idea of “craft” in design. Machines, he says, reduce the control a person has over the process of creating an object. Because they are rigidly designed to produce the same thing over and over again, the versatility inherent in the design process is decreased. “And as versatility decreases,” Osborne writes, “the demand for human skill diminishes, or at least changes direction.” The decoupling of design and production (or the segmenting of design into its own, smaller space of production; that is, production for an audience that only includes the co-producers, engineers and operators) sharpens this turn. Osborne says that, when design is sectioned off this way, it becomes “an ideal to which the product approximates according to the suitability of the machine and the quality of the workmanship.”
Sub in “machine” for either Figma or the operating system, and “workmanship” for engineering, product, and all the other various functions with which we operate every day as interface designers. Osborne was writing this in 1977, but his framing of our situation feels like it was written yesterday.
What’s different, of course, is that, today, we are invoking the idea of “craft” in opposition to a new type of automation that is actually valued for its variability rather than a consistent and perfect mode of reproduction. Language-driven AI models are unpredictable, provide varied results, and—right now—can’t seem to keep track of things very well, particularly visual and aesthetic signals. These are characteristics that should exclude the technology from the exaltation enjoyed by machines of mass-production. And if they lie in opposition to the accuracy of those machines, maybe they should increase rather than diminish the demand for human craft—should enable a new coalescence around human determination in the process of design. But, so far, it seems like they don’t.
This could be because unpredictable behavior subconsciously convinces us a machine or model is alive, as Judith Donath describes in The Social Machine:
When machines work exactly as we expect them to and do what we request of them, we think of them as simply machines. It is when they do not work as expected that they appear to have a will of their own and we ascribe intelligence to them.
Or, it may simply be because this unpredictable nature signals to us that there is a wide field of possible outputs and that we must merely narrow down the use of the machine (Osborne’s “quality of workmanship”) in order to use it to its best potential. But our interaction, so far, mostly takes place through human language which—besides being an inherently subjective mode of expression, ultimately impervious to “engineering”—is imprecise, interpreted, and indexical in meaning.
But neither of these potential understandings of AI in design helps us define the “craft” of design and where it fits into the future.
Craft, in Osborne’s view is concerned with what I see as a few central themes: control, continuity, and rigor.
Control
In our context, control means two things. First, we could think of it as direct control over our tools: having the tools, recognizing what they are, and understanding how to manipulate them in a way that transforms our base materials into a new object in the way we intend.
Of course, almost all of the tools we use on a daily basis are digital. They are mechanical. They therefore place a limitation on this portion of our craft, having already abstracted away much of the human skill and expression we can exercise. On a side note, this is precisely why I’m becoming increasingly radicalized (again) toward the notebook and pencil as tools of creative exploration.
Second, control can mean control throughout the process of production. Involvement in all steps of production; the constant “friction” we so often describe between designer and developer being resolved by a deep cooperation in each others’ practices. This doesn’t mean that a specialized designer should unilaterally make every decision throughout the process, just that each practitioner’s hands should be working at every step. Most often, this looks like iteration based on technical feedback, product direction, etc. Craft is made collective, and is strengthened by constant, deep involvement in each others’ work, a mode that is often disincentivized by specialization and demands for increased productivity.
But regardless, behind every designer that we would think of as one of the greats lies not an empty white room but a team of collaborators and equals who all touched the final product. A collective hand wielding a complex, multifaceted set of tools.
Continuity
Osborne refers often to crafts (like stonework or iron smithing) as having a long continuity of experience in terms of how long the craft itself has existed. Most of the time, he mentions crafts that have been around for hundreds of years. He does this to invoke the idea of acquired knowledge of tools, colloquial techniques and histories one absorbs by learning from their creative forebears. Also relevant is the notion of transforming base materials into new objects; environmental or world transformation through creative acts.
This, I think, is interesting when it comes to the software interface. When it comes to human-computer interaction, we just don’t have hundreds of years to refer to. A lot of our practice’s acquired knowledge comes from relatively recent developments and research, and is supplemented by knowledge from fields like graphic design, architecture, and—although less than is necessary—the social sciences. This supplementary heritage is part of the reason I started Design Notes—to prove that there’s a lot to learn from other creative fields, other tools, and other colloquial histories.
The other thing that interests me about this principle is that learning to use tools, particularly complex tools that require advanced physical manipulation in addition to intellectual rigor, require social learning. Anthropologists observe that tools reach a threshold of complexity beyond which no individual could create the tool in their lifetime, meaning that a collective approach with collaborative learning is needed to move forward with the tools at hand.
Rigor
I think this is the kind of “water in the sponge” of the two other principles of craft that I’ve extrapolated from Osborne’s work. Rigor is a vague term, but Osborne makes the idea concrete by describing it as a “cult of perfection” that leads craftspeople to achieve the subjective and meaningful feeling we all experience when we encounter a well-crafted work. Osborne cites himself in the Oxford Companion to the Decorative Arts here:
[Craftsmanship] involves a genuine pride in the process of production itself, a pride which drives a man to make whatever things he makes as well as they can be made, even beyond economic considerations of reward.
In the frame of this post, rigor is the depth of the work toward this pride.
In my practice as a designer, I see this manifest most often as the rationale of the work. Our work designing the interface is systematic, and the systems we construct touch numerous other systems—in the corporate environment, in the overall cohesion of software, and in people’s actual lives—and the extent to which we are aware of, acknowledge, and deal with that in the work is our rigor. That comes down to making well-informed decisions and orienting our expert use of our tools toward making them.
Osborne’s point about economic reward can’t go ignored in this conversation: as I’ve written before, there is a strong inclination toward AI as a tool of productivity; explicitly toward optimizing economic reward from the production of material culture. And I think this is one exposed point in which these principles of craft can and must enter the conversation.
It is, ultimately, the human motivation to create, to create well, and to exercise and extend our capacities of rationale, tool use, and their junction—craft—that will push design forward.
In any approach to integrating automation with the interface, these principles might have a chance of guiding us toward a well-crafted future.
