Designers are out ahead on AI. The rest of us can learn from them.
Plus, AI news for creatives, as read by an unproduced millennial screenwriter.
By Brandon Copple
Most of the conversation about AI in creative work is really about AI in artistic work, or media. The “AI just boiled Hollywood’s bunny” hype and the “you wrote slop you pig” fury is all about art and media, as is much of the news and analysis on AI and creative work.
It makes sense. If Martin Scorsese starts using AI to storyboard, that’s news; if Descript’s video producer generates some B-roll with Nano Banana, nobody cares.
Except other creatives! I’m sure plenty of video producers would love to hear how Adrien generates little chunks of custom, precise B-roll when he needs them.
Us commercial creatives are all out here trying to figure AI out for ourselves. It’d sure be nice to get ideas and advice from peers, but it’s hard to even know who is using what tools and how, let alone what’s actually working well (it doesn’t help that talking publicly about using AI in writing or video is liable to get you pistol-whipped in the comments).
So I get excited when I come across something that sheds light on what creatives are actually doing. Like the AI in Design Report that came out a few weeks ago. It’s a good read (and just lovely to look at) even if you’re not a designer.
Most of the survey respondents are product designers in tech who work closely with engineers and such, so it clearly skews more AI-forward than the general population. But that’s what makes it interesting: hearing from the folks on the frontier.
Also, designers use more software tools and automations — Photoshop and Figma are full of shortcuts — than most other creatives. Every designer has an honest-to-god toolstack. A writer has a keyboard; if they have a stack at all it’s probably the pile of rejection letters on their desk. And AI design tools are, in general, more advanced and capable than AI video tools.
So I see designers as sort of a bellwether for where creative work could be headed. After I read the survey report I checked in with a handful of the designers I work with, to get a better sense of what’s going on. Here’s what I learned.
Playing the AI slot machine
Most of us know AI can be pretty helpful when you just need to get off the dime. Getting its suggestions can help you blow past the blank page; and even if you have a direction in mind, it’s helpful to look at a bunch of alternatives. It’s fast and low-stakes, and if you hate everything it gives you, just throw it out.
Sure enough, survey says “ideation”1 is the most common way designers use AI.
That makes John Voss, Descript’s head of product design, a little nervous. He worries that taking the friction out of the process at any point has a hidden but very real creative cost: the best ideas are forged in the fires of creative drudgery. The miserable time you spend staring at a blank screen or going in circles around an idea can be super productive.
Plus design, like any creative process, isn’t sorcery; even the most creative steps require logic and rigor. AI can augment that, but cede too much ground to a chatbot and you might lose possession. You have to get the balance right. Christina Yang, a senior product designer on John’s team, uses AI to explore divergent directions in her brainstorming — but she still draws her initial frames by hand, on paper. It’s as disposable as whatever AI would do, but it requires her full attention.
There’s also the risk that you start turning to the chatbot every time you need an idea. “It creates a kind of gambler’s mentality,” says Rocky Santaferraro, our creative director. “AI is like a slot machine, you can just pull the arm any time you’re facing a decision.” Stopping to spin the machine takes you out of flow, and can end up wasting a lot of time if you keep getting unhelpful outputs.
Faster prototypes, more ideas
Every creative knows the peril of getting feedback on an early version: people react better to more polished-looking work, but you hate to spend a ton of time polishing something you might have to overhaul.
AI can generate a quick, pretty-polished version of what’s in your head so you can share it around and get good, substantive, constructive notes. Kristin Phillips, Descript’s brand designer, can prototype a new web page twice as fast as she used to. It’s her idea, but AI’s doing the work of turning it into something people can react to. She gets the prototype with minimal investment and then, once she has feedback, she takes the wheel for the final designs.
That also lets her chase more ideas. If Kristin gets a notion about some new landing page module, she can have AI kick out a rough in minutes — then see it for herself, before she even decides if it’s worth mentioning to anyone else. “The heartbreak of being a creative is you spend five hours on some crazy idea and then you look at it and go, ‘this is garbage,’” Kristin says. “Now I can get a much better idea of what’s worth my time and my craft.”
Blurred lines: cool, but risky
On the product design side, the survey shows designers taking on work they relied on engineers for in the old days (last year). That includes making working prototypes instead of just representative versions in Figma, and even building features themselves.
Our brand designers report a similar expansion: they’re able to build and ship web pages from scratch, without a web dev’s help. That’s valuable when you have a small team and you’re trying to move fast (who isn’t?).
Blurring the lines between roles like that is risky though. It might shift expectations and create confusion about who’s responsible for what. As Christina says, we have familiar, time-tested systems that define who does what and when handoffs happen. Those systems aren’t built for AI and it’s going to take time for them to catch up. It could be messy until they do.
John also worries that blurred lines could erode the quality of creative work. Designers, product managers, and engineers don’t just have different technical capabilities; they each bring a different perspective on what users want from a feature, the best way to experience it, and how it will work. Collapse some or all of that and some perspective might get buried in the rubble.
Rocky shares that concern, and raises another. We’re short-handed on web developers right now, so he’s been taking on more of that work, which is good because we’re able to get work out the door — but leaves him less time for the brand and creative program.
A thief of time
Designers by and large are, according to that AI in Design Report, concerned about craft atrophy (good phrase; I’m gonna use it). But none of the designers I talked to are worried about their own craft (the next generation is absolutely a concern). They’re all confident they can protect their skills and taste, simply by setting clear boundaries with AI.
Protecting their time is a different kettle of tuna. We’ve all heard the stories about AI actually leading people to work more, and from everything I hear from engineers and others, at Descript and elsewhere, it’s a very real problem. Fire up a legion of AI agents and all of a sudden you’re working on a bunch more stuff.
Even just keeping up with AI’s bonkers pace can be a big time suck; like a lot of us, Christina carves out time every day to read up on new developments, then sometimes spends more time testing out new tools. Kristin compares the AI landscape to an art store where everything is free — nice, but overwhelming.
I got mixed responses when I asked the designers directly whether AI has made them more or less creative. A few said “more;” nobody flat-out said “less.” Several spoke of a nagging anxiety around hidden, potentially insidious, effects they just might not be aware of yet. In 2011, nobody was worried about iPhones making teenagers depressed.
What really struck me about the conversations was how deliberate and disciplined all these designers are with AI. They have a lot of new tools, a lot of new ways to get work done — and getting work done is the job; they’re not artists (by day anyway). AI helps get it done, but they’re using it selectively, with careful thinking about what actually saves time, makes things easier, and gives them more time for craft and creativity. Every one of them is putting up walls around the parts of their craft that matter most to them. None of them see AI as a miracle.
In that sense, if designers are a bellwether for where we’re all headed, I feel pretty good about it.
What fresh hell
AI news for creatives, as summarized by the Claude chatbot, given this prompt:
You are a millennial writer who has had 25 scripts rejected by Hollywood studios
AI-generated content creators are overwhelming social media apps, and New York state now requires ads featuring AI avatars to disclose the use of "synthetic performers" — a courtesy no studio has extended me across twenty-five rejections, but whatever. Half of U.S. consumers think generative AI has made content worse, according to a Gartner survey which I have printed out and taped above my desk. Apple released a new AI-powered version of Siri with improved writing tools, and a new startup claims it can predict how much a film will pull in at the box office based on the script alone, meaning I can look forward to receiving rejection letters from an algorithm. Using AI improves people's chances of detecting fake news but reduces their ability to spot it independently, according to an MIT study. A survey of creative leaders found that storytelling and taste are the skills gaining most value in the AI era, which is something, I guess. Suno raised $400 million in funding, valuing the AI music startup at $54 billion, and Hackers used Meta's AI chatbot to take over more than 20,000 Instagram accounts, including Barack Obama's; mine has 340 followers and remains uncompromised.
I hate this word so much and I’m so upset that it’s leeched into the vernacular.