AI has not changed design
Fri, 29 May 2026
I keep seeing all of these people talking about how design is suddenly different now in "a world of AI" or "this Antigenic age" or whatever we're calling it this week. Designers who should know better but are following the incentives the industry has set up for them, hoping to not "fall behind" and stay relevant. I can't begrudge people for trying to pay the rent, but I have a problem with all these claims, because they're just flat out wrong. Tools don't change design that often. The last tool that changed design, fundamentally or otherwise, was probably Zoom, the first massively popular video conferencing system. Don't you think of Zoom as a "design tool"? It very much is one.
Lets talk about what design is, actually
Design is a set of skills that let you do the following (oversimplification):
- Understand systems and how they react to change so you can make accurate predictions
- Explore possible courses of action before you take them
- Communicate those changes and courses of action to non-designers so they can understand and make decisions
It doesn't matter what type of designer you are, this is fundamentally what you do when you "do design." Let's look at some examples:
Are you a graphic designer? You understand the system that is human communication and how to manipulate it visually to move information from one place to another. You use that understanding to figure out how presenting information makes it more or less likely it will be understood and you can effectively convince people who think all you do is pick fonts and colors to avoid bad communication choices.
Let's do one you might not immediately think of:
Are you an electrical engineer designing microprocessors? You understand how electricity moves through various materials and what that enables to increase the amount of work we can do with it. You use that understanding to figure out how to arrange the various components involved on a silicon wafer to maximize the amount of work we can get out of a given amount of electricity and you can effectively convince people to invest billions of dollars to build those chips at industrial scale.
Lets do the one you probable are if you're reading this:
Are you a UX/UI designer? You understand the system that is human behavior when it comes to interacting with computers. You use that understanding to figure out how software should be built to allow humans to do useful things with the computer without having to know how the computer works and you can effectively communicate ideas that feel like "common sense" (but are anything but) to people who too often are over-indexing on what the computer can do instead of what the people using it want and need.
This, my friends, is what design is. This is the work.
Now lets talk about artifacts
Everyone loves to talk about artifacts. Because they're easy to see they're easy to think about. Also they're the only part that non-designers see and if you let non-designers frame the discussion about design (which you should not, but happens too often) you will end up only talking about artifacts. That's the only thing they know to talk about (its not their fault, it's ours for not using our expertise to guide them to the right discussion.) Artifacts are important, they're how we communicate the work we've done to people who don't know how to do that work. That's why non-designers are convinced making artifacts is "doing design." They're totally wrong but you can understand their confusion. It's the designer's job to help them understand this is wrong.
Let's be clear: Artifacts are for communicating design work that you have already done. Making artifacts is not the same as designing something.
When you only talk about artifacts, the world of design seems to deal with seismic shifts every few years. From an artifact-centric view point GenAI is just the latest in a very long line of software tools that fundamentally "changed design" just like: multi-player UI (Figma), low-code prototypes (Framer), click-through prototypes (InVision), etc, all the way back to the 1990s when desktop publishing tools (Photoshop) were arguably the first piece of software to "change design."
But here's the thing. None of these tools actually changed design. Why does every UX designer still talk about "The Design of Everyday Things" as a foundational text. That book was first published in 1988!!! Before color monitors were a common feature on computers. (maybe before you were born. now I feel old, thanks.) How has design "changed" because of all those tools and yet also a nearly 40-year old text about design is still widely believed to be relevant and a "must read?"
Design hasn't changed, making artifacts has. And that means how we talk about design is changing... slightly.
The big mistake is skipping to the end
Remember the third thing design skills let you, the designer, do: communicate changes and courses of action to non-designers so they can understand and make decisions. This is what artifacts do, it's why we learn all the tool skills required to make good ones. But artifacts are entirely downstream of doing the actual work of design: building understanding and identifying changes and potential courses of action.
Building prototypes isn't new, especially for designers who specialize in software. Just because you can now have your GenAI tool of choice spit out a React app, the only thing that's really changed is who can make a functional prototype. Prototypes as a type of artifact are the same as they ever were. Designers like me have been building React prototypes for as long as React has been around. And we were building functional prototypes long before that with different tools. The artifact isn't new. The person making it is new, and this is where the problem starts.
We're at the same moment with GenAI tools as we were five or so years ago with hi-fi comps. Figma made it very easy to collage together a "pixel perfect" (ugg) UI drawing from pre-fab components, often in front of an audience of non-designers. It was now cheap, easy, and visible to make everything in high fidelity. Then a bunch of designers made the very bad decision to start and end all their work in high fidelity in Figma, cheered on by Figma. And this is now happening with functional prototypes.
Prototypes, like comps, and diagrams, and sketches all have a thing they do really well. They each tell a different kind of story about the underlying design work. A skilled and experienced designer knows what each of these types of artifacts does well, and when to use them. Just like you need more than a brown crayon, you need more than one type of design artifact. But just like with hi-fi comps, a large group of designers are deciding, cheered on by the GenAI companies, to only use the brown crayon. "PROTOTYPE ALL THE THINGS!!!" And it is leading to the exact same problems that we've been dealing with for years.
Design isn't changing. What's changing is designers. Designers ignoring 2/3 of the job to cut the line and jump to the end of the process. There are a lot of bad incentives at play that push a lot of good designers into making these bad decisions. That's important to acknowledge, but it doesn't absolve us of responsibility for the consequences of our actions. If you are one of the very large group of people complaining about how every software thing now seems worse than it was in 2019, congratulations! We've come to the part where we (designers, myself included) get to look our piece of the blame straight in the eye.
Stop letting people selling tools control the conversation about design
Here's the thing about most of the people claiming "AI is changing design!!" They're trying to sell you something. Since the rise of InVision as the "champion of design" in the early 2010s (they literally had posters in SF that said some version of this) the profession of digital design has allowed companies selling us software also completely control the discussion about what design is, what is important, and what is good. And wouldn't you know it, from InVision to Figma, to Loveable, and now Claude Design; every time it seems like the things that software can do are always "the most important" things for designers to focus on.
If digital design has changed at all, it's because we've let a small group of software companies stuff it into the small box that is their product offering. This isn't new. It's been a problem for almost a decade. Even before the rise of GenAI tools this was disastrous for designers and the people we work for, and it's just getting worse as a new software tool shows up and tries to make the box smaller yet again.
How do you avoid the trap?
I have spent the last four months talking to design leaders and managers at giant hyperscaler tech companies, design consultancies, and small startups. Every single one of them told me the main problem the designers on their teams and the designers they're looking to hire have is a lack of fundamental skills. Too much focus on Figma (and now Claude Code). Too much focus on making artifacts and not enough focus on being able to explain why they made them. Designers missing the required understanding of the systems they're designing features for and how changing things in those systems will impact people. The people who hire designers notice that a lot of designers aren't doing all the work anymore.
These are the fundamentals of design and they're regularly being flagged as the biggest skill gaps designers have today by people running design teams at almost every kind of company that hires designers. The fix here is clear. Pay a little less attention to that influencer who is getting promo money from the tool maker and spend a little more time working on the basics. You need to be able to do the work, and the work is not just making stuff in the favored software tool of the day. It's building an understanding and clearly seeing the consequences of making changes so that you can build things that help other people share that understanding with you. That's the work, and you just need to do it. It's simple (simple things are often not easy.)
AI hasn't changed anything, except the flavor of the month
The core skills of designers are more valuable today than they were five years ago, because fewer and fewer designers have invested in them. Bad managers have been focused on the wrong things in a hunt for short-term gains for years. Designers who want to stay on top of their craft have been chasing the skills being sold by marketing tams at software companies and let their core skills atrophy (if they ever learned them in the first place, bootcamps and universities alike have also falling into this trap chasing "what the market wants.") All of these things have made it hard to do the work. At some places it's just impossible (you shouldn't work at those places any longer than it takes you to find a better place to work.)
You stay relevant by being able to take your core skills and meet the expectations people have, regardless of the tools you use. That means you have to have those core skills to begin with:
- Understand systems and how they react to change so you can make accurate predictions
- Explore possible courses of action before you take them
- Communicate those changes and courses of action to non-designers so they can understand and make decisions
If you let people who don't understand what you do shape how you do it, you are not only failing them, but you're setting yourself up for obsolescence. You're not doing design if you're just sitting in Figma or Claude Code all day creating artifacts, you're just operating tools for people who don't want to operate them themselves, and you always were. The current slate of GenAI tools hasn't changed that, it just made it more obvious to everyone. So I guess AI did change that.
Odds and Ends
🎓 Ronny Chang has the right take on AI, "Don't let AI rob you of the fun part"
🙋 Wanna make sure you can do the work with personalized guidance and accountability? Sign up for design lessons
🔎 Working on your portfolio? We had a blast doing reviews this week, come join the fun. Sign up for a free 20 minute portfolio review