Faster, not better?

Last month Full Moon released our summer research on how AI is changing design. We then held a webinar with five great design leaders to discuss our conclusions. The panel was Merritt Robinson (Global Head of Design and Research, Bain & Company), Mat Hunter (CEO, Design Council), Roger Rohatgi (VP UX, The Home Depot), Olof Schybergson (co-founder, Fjord) and Shelley Evenson (former head of design learning, Fjord/Accenture). Here (at least for us) are some of the most striking things we heard. Some have been lightly edited for easier comprehension.

One word to describe how you feel about this topic

We opened as we opened every one of conversations in the research, by asking for one word that sums up how they feel about AI and design right now. Mat chose anxious. Part of that is anxiety for practitioners, and part is anxiety for the profession's voice: "I'm anxious that we will be seen as invisible and irrelevant." Merritt chose anticipatory: "These are amazing tools, but what do I do with them?" Roger chose symbiotic. Olof chose energized: "It can be really tough and disastrous, but it's also an opportunity to rethink, to retool." Shelley would pick “conflicted”. She lives two blocks from OpenAI, and "the story is so confusing." (this was just after the exposure of the now famous Hugging Face incident). This really mirrors the range of what we heard in the research. Pragmatic acceptance, excitement, nervousness.

Faster, not better

Nobody in our research said AI had made their work better. Mat questioned the premise first: "Define the metric." He also put the tools themselves in context, drawing on his years building digital tools in the 1990s:

"A violin is a very unforgiving instrument... A six-string guitar is a little bit more inclusive, in that you can do a pretty good job with it in an hour and then a really good job in a lifetime. And then a sort of automatic Bontempi organ... that's highly democratic in terms of it being accessible, but bloody awful for a really good musician." His verdict: "We are in the foothills of it."

Merritt made an important distinction. The gains vary by task, not by person. On research: "That sheer access to the user, that ability to do that at scale, to me is just an incredible accelerant." On journey maps: "If you try and use all the new tools to do journey maps right now, it is harder, longer, worse... sometimes we’re saying, let’s go back to what we know until we get the new thing worked out."

Working with intent

One research participant said that "the more you know what you want, the harder it is to work with it." Shelley agreed, and described asking Claude how she should collaborate with it. The answer was that she should be more precise. "That's exactly what I don't want. What I want to do is explore and experiment and see what's going to come of the conversation." Her teaching experience points the same way: "It's really important not to overprescribe."

On another occasion she asked for an interface that wasn't a chatbox. "It produced what I wanted, but added a chat box at the bottom of it. Just in case." That – to us - feels like a classic example of AI forcing design and content to the middle of the road.

The language layer

Olof reframed a major point from the research about language as a new layer between designer and design. A cursor and a stylus were intermediaries too, so language replaces a layer rather than adding one. What is new is whose interface it is: "That is our UI with each other. So suddenly AI is using our UI."

Merritt described what gets lost in the switch. Teaching her stepson to draw, she found she couldn't explain how an idea gets from your head to the page. Language "loses that connective tissue where you're working things out and the idea is coming to life."

Where the speed went

Roger set out a structural problem in working with AI which may will recognise: "There's a major delta between how fast we can move versus how fast we can judge." He also offered a warning he calls AX, or artificial experience: "AI designing for humans on what AI thinks humans need or want." His conclusion: "The biggest decision we'll make as designers, as a company, or as humanity, is deciding what decisions we want AI to decide."

Merritt both challenged what has changed and gave a old solution for overload a new frame: "My calendar was packed before AI and it's packed now." Her team protects a ritual it calls answer time. "It is not bot time. It is human time."

Don't blame the new tools

On the flattening of design, Olof was blunt: "We can't blame AI for the flattening of design. It's been going on for a long period of time. Many of us design practitioners got very well remunerated doing pretty repetitive and tedious things in digital... AI is actually quite good at that." He sees the "more with fewer people" story as a quality problem and trap: "We assume that we're going to do the same old crap just faster with fewer people."

Mat raised the economic backdrop. Design can still read as "slightly shiny, slightly self-congratulatory, definitely slightly conspicuous consumption." Olof described the mood of the times: "Companies, organisations, politicians, they're just in a hard place at the moment." He expects the pendulum to swing back, though: "A world without designers is not a good world."

The juniors problem is real

Merritt gave the most honest answer of the session: "My senior folks who are phenomenal, we got them when they were junior, and we grew them and we apprenticed them... I don't know how you get to that without that apprenticeship."

Who holds the pencil

Neither Merritt nor Roger saw non-designers designing as new. Roger pointed to Photoshop and Figma files from business analysts. Merritt said the pattern usually ends the same way: "It's not good yet, and one way or another it's getting punted back to the design team." Her advice is to hand the tools over "with some generosity and pride."

Of our four possible positions for design, Shelley chose accountability for the human at the other end. She argues for a "human AI experience officer." Merritt grounded it: "There's a human as a human and a human as a customer... most of us don't get to be designers unless there's a business part of it, otherwise we're artists."

Mat drew a line between two ways of knowing what works. One is Google testing hyperlink colours empirically, which he called "a real sort of Darwinian thing." The other is judgment: "the wisdom to predict whether something will succeed or fail because you can't afford the risk of it going wrong."

Olof, crediting Fabio Sergio, predicted where value will move. There will be more in craft and quality and a resurgence in strategy, and less in the industrialisation of digital: "There will still be roles there, but not as many."

Roger had the last word: "We are not prompt pushers, pixel pushers. We are the imaginators."

From the chat

Edwin offered a wonderful line: “imagination is the capability that bails out the myopia of logic.” Thanks Edwin for that! Several others raised accountability – in our view a topic that will get hotter and hotter, and not just for designers. We can't hold a model accountable for what we make with it. If it doesn't work, that's down to us.