Why designers use AI and still recoil

The claim

Designers recoil from AI, the usual account goes, because it makes poor work or because they fear for their jobs. Two 2026 surveys look at the question from opposite sides, one of architecture practices that have adopted AI and one of artists who mostly have not.

Architects: using it, and not crediting it

The Royal Institute of British Architects published its third member survey on AI in July 2026. The share of practices using AI for most projects rose from 59% in 2025 to 74%. 75% report better productivity and 57% a positive return on investment. Only 17% agree their designs are better because of AI. 77% agree it can never replace human creativity, and only 21% say it makes them more optimistic about their organisation’s future. 59% expect it to lead to staff reductions across the profession, and 61% say it will be harder for early-career staff to gain skills and experience, because AI is taking the tasks they learned on. The release does not give the number of respondents.

So adoption is for the work around the design. The practices that use it most do not say it improves the design.

Artists: refusing it

A team at a US university surveyed 378 professional visual artists, among them illustrators, graphic designers, videographers and tattoo artists. 85% never use generative AI in their work and 88% never use image generators. 99% dislike it and 92% dislike it strongly. 45% meet AI-generated images in their practice daily. 80% believe they compete with it, 90% say it has reduced their income opportunities, and 61% report disruption to their work or business.

The respondents were recruited through flyers at an artists’ convention, trusted artists, and social media, and were asked for a portfolio as proof. Most live in the United States, Canada or Europe, and about two thirds identified as LGBTQ+. The 99% describes these artists, not designers in general.

Their reasons are mixed. Some tried early tools and found they “consistently failed to produce useable results”. Others stopped on learning the models drew on artists’ work without permission: “When I learned they were sourcing their art without permission of the artist, I stopped using them.” Those who do use it often do so under pressure from clients, bosses or the tools they already work in. And some described a new burden. One illustrator says people online look at art “less to enjoy it and more so to figure out if its AI generated or not”, and that there is “pressure and anxiety in proving you are a real person now.”

A separate interview study of 15 queer artists finds a tension between how their art is made, through community and relation, and what they see as the anti-relational way AI is built and used. Refusal is their response.

What the two surveys share

Quality alone does not explain either. The architects use tools they do not credit with better design. The artists who refuse give quality as one reason among several, next to consent and competition. If quality were the cause, rising quality would end the objection. The artists’ own accounts suggest it would end only the first reason.

Both groups worry about the people at the start of the profession. The architects’ concern is that junior staff lose the tasks that teach them. The artists report fewer opportunities, and one describes younger artists giving up.

Whose reasoning made it

A designer’s work is judged partly by whose it is. Each of these objections turns on authorship: whose work trained the model, who is credited, who has to prove they made a thing. A tool that produces a better picture addresses none of them.

What it means for our work

The artists’ objections split in two. Some concern how the models were built, whose work trained them and who now competes with them, and a framework changes neither. Others concern what a tool does with a person’s work and name, and who is credited for the result.

The second kind is where the lab’s rules apply. Ideas are made from a named framework, a written account of how one person reasons, so the question “whose reasoning is this?” has an answer on the page. A framework of a living person comes only from their public words, is labelled as our reading, never speaks in the first person, and comes down if they ask (What personal AI can cost).

The architects’ first worry, early-career skills, bears on the same point. Early ideation is the work junior designers learn on. A framework gives a junior designer the reasoning of a master in a form they can read, argue with and reject. An image generator hands back a finished picture and no reasoning. Where ideation is how juniors learn, the first leaves the learning in their hands.