Generative tools save valuable time, but they also end up doing the design for us. The question remains: which tasks can we delegate to them without harming our expertise , and which ones, however tedious and time-consuming , are those that shape our development?

I attended the Figma Config 2026 keynote . The announcements kept coming: shaders generated from a simple description, a true animation timeline integrated into the canvas, plugins created by defining an intent, an agent capable of executing entire workflows. I was blown away. I could already envision the applications: time saved, more compelling deliverables, and better-engaged clients.

And then, it hit me . With a prompt, Figma did what we were doing at the very beginning of our career.

I also saw myself at 12 years old, the day I downloaded Photoshop. I learned on my own, by trial and error: retouching a photo, designing a poster, animating, searching for hours for the right effect, the right typography, the right texture, learning to work with layers and masks. I especially remember the joy that came afterward, when, after several failed attempts, it finally worked. And often, it was by making mistakes that I discovered something else: an unexpected process, a result I hadn't sought and that I ended up preferring. Those hours of creation, mistakes included, weren't just technical training. It's where my eye , my technique , my designer's logic were formed.

Today, someone starting out doesn't necessarily need to go through that process. They describe what they want—an effect, an animation, a product, an interface—and the tool creates it for them. The time saved is real. But I ask myself a simple question: if we no longer create things ourselves, how do we still learn? And ultimately, do you really need to know how to do things to become a good designer?

Judging and acting: distinction and dependence

An experienced designer who rejects a project doesn't always touch the file: they see that a typographic hierarchy is weak, that a volume lacks tension, that a navigation isn't intuitive. They see what needs correcting, even if they can no longer always implement the correction themselves. But their eye was trained while they were still executing. Judging and creating are two different skills. The developer who reviews generated code in a few seconds is in the same position: they judge without writing, but they can only do so because they wrote beforehand.

Because this eye , even when it seems like a gift, is acquired . It is formed by conceiving, failing, and arbitrating, project after project, for years. This is Dewey's entire thesis: we don't learn by receiving, we learn by doing, and then by experiencing the consequences of our actions. Bourdieu calls this embodied flair , this "sense of the game" that no rule transmits and that only prolonged immersion establishes, habitus. Rumi, long before them, already contrasted knowledge received from the outside, which he compared to water brought in through a pipe, and which eventually stagnates, with knowledge verified by oneself, which springs forth like a source and remains vibrant. Munari, for his part, reminds us of the maker's perspective: imagination doesn't emerge from a void, it grows with what one masters. Conceiving becomes easy, he said, when you know how to do it.

Not all time-consuming tasks are created equal.

This is where the real distinction lies. Some tasks are tedious and disposable : cropping, resizing a hundred visuals, exporting variations, renaming layers, updating a design system across thirty screens, wiring navigation links in a prototype. We understand these from the first screen; the next twenty-nine teach us nothing. Automating them adds nothing to our training.

Others are arduous because they shape us : composing a layout from scratch, iterating on a product form, designing a path by exploring multiple routes, finding the right hierarchy of information, and above all, successfully executing a complex transition or effect: searching, making mistakes, starting again, until we understand how it works. These "lost" hours build our eye, our technique, and gradually the speed with which we execute them. Rushing through them means obtaining the result without understanding or mastery.

The criterion depends on one's experience level: a task that is valuable training for a beginner becomes disposable for a senior employee who has done it a thousand times. This is the logic and responsibility of management: a good manager assigns a thankless task to a junior employee not for efficiency, but because that's where they grow. If AI absorbs these training tasks, the junior employee loses their training ground . They don't produce or judge enough to develop their skills or judgment. Freire already contrasted the passive deposit of knowledge, where information is poured into a mind that receives it without acting, with praxis , where action and reflection nourish each other. Morin adds that judgment is a faculty that is trained: a well-formed mind, capable of connecting ideas and identifying errors, is not built by accumulating ready-made answers. A recent survey conducted by Microsoft Research and Carnegie Mellon supports this idea: professionals who maintain a critical eye on AI productions are those who have confidence in their own skills.

Without playing the prophet of extinction

This isn't about announcing the end of everything, as when people thought printing would kill writing or photography would kill painting. Each time, the old has survived, and often gained new value. The work of the designer already demonstrates this: when production becomes abundant and accessible, value shifts elsewhere.

The gesture , first of all, doesn't disappear: it changes status . Baudrillard illuminates this: in a world saturated with copies without originals, what becomes rare becomes precious. The handmade, the visible process , the trace of trial and error regain value precisely because they have become the exception. An interface generated in three seconds is barely distinguishable from a thousand others; what makes it unique is the ground that was explored, the users that were observed, the prototypes that were discarded before this one. What was thought lost returns through its value.

Impact then takes over. As Papanek reminded us half a century ago, a design is not judged by the allure of its form, but by its ecological, social, and political consequences . What will distinguish one work from another tomorrow has less to do with the quality of the surface than with the questions that preceded it. For whom? At what cost? With what care? As the machine absorbs the execution, what remains for the designer is precisely what Papanek identified as the essence of the profession: the responsibility for what one brings into the world.

What to do

The first task of the designer, and especially the manager, is now to prioritize . Which tasks should be delegated to the machine without losing anything? Which should be kept, not out of nostalgia for the physical act, but because they shape the person performing them? Automating a tedious task without considering what it taught the junior employee is to mistake saving time for progress.

Findeli offers the criterion: in design, one only truly thinks in action , within a real project. Protecting training, therefore, does not mean banning tools. It means preserving, in each learning path, the moments when one designs from start to finish often enough, for long enough, so that the hand , the eye , and the mind can develop together.

So perhaps the real question is not "do we need to know how to do it?", but the one we should ask ourselves before each task we leave to the machine: is it time we save , or learning we lose?

References

• Baudrillard, Jean. Simulacra and simulationParis, Galilée, 1981.
• Bourdieu, Pierre. PracticalityParis, Éditions de Minuit, 1980.
• Dewey, John. Art as an experience. Trad. J.-P. Cometti et al., Paris, Gallimard, coll. « Folio Essais », 2010 [ed. orig. Art as Experience, 1934].
• Findeli, Alain. “Project-based research: a method for design research”. Proceedings of the first Design Research Symposium, Swiss Design Network, Basel, 2005.
• Freire, Paulo. Pedagogy of the oppressed. Paris, La Découverte, 2001 [original edition] Pedagogy of the Oppressed, 1968].
• Lee, Hao-Ping (Hank), Advait Sarkar, Lev Tankelevitch, Ian Drosos, Sean Rintel, Richard Banks and Nicholas Wilson. “The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers.” Proceedings of the 2025 CHI Conference on Human Factors in Computing SystemsYokohama, ACM, 2025, pp. 1-22. DOI: 10.1145/3706598.3713778. Open access PDF: https://www.microsoft.com/en-us/research/wp-content/uploads/2025/01/lee_2025_ai_critical_thinking_survey.pdf
• Morin, Edgar. The Well-Formed Mind. Rethinking reform, reforming thought. Paris, Seuil, 1999.
• Munari, Bruno. Da cosa nace cosa. Supported by a software methodologyRome-Bari, Laterza, 1981.
• Papanek, Victor. Design for the Real World: Human Ecology and Social Change. New York, Pantheon Books, 1971.
• Rumi, Jalal ad-Din. Mathnawi, book IV. 13th century.

 

Hend Daoud Metoui, Product Designer at UX-Republic