There is a familiar argument whenever generative AI enters a conversation about creative work: artists hate it, musicians hate it, writers hate it, and everyone is worried about being replaced.
But the backlash is not necessarily the same across creative industries.
Recent research from the UK’s Association of Illustrators (AOI) suggests that illustrators are experiencing a particularly direct form of disruption. In its latest survey, 46% of respondents said they knew they had lost work because a client chose generative AI instead. That figure was 32% in the previous survey. The latest survey received 549 responses.
That distinction matters. The argument is no longer simply that AI might threaten artists in the future. For a substantial proportion of illustrators surveyed, the technology is already appearing in the place where it matters most: the commission.
The difference between being replaced and being asked to compete with a machine
Illustration has always been a commercial service.
An illustrator is hired to create a book cover, character design, advertising image, concept art, editorial illustration or visual development. A client has a problem, and the illustrator is paid to solve it.
Generative image models have entered that exact transaction.
The AOI survey found that 62% of respondents had encountered AI-generated images in briefs or moodboards. Meanwhile, 36% said they had been asked to copy, refine or redraw AI-generated imagery. Another 31% reported pressure from clients to accept lower fees, while 27% reported pressure for faster turnaround times.
That creates a particularly uncomfortable situation for illustrators.
The technology is not merely producing something that looks vaguely like their work. It can become part of the client’s expectation before the illustrator has even been hired.
A client can generate an image, decide that it is the direction they want, and then approach an illustrator with something close to a finished visual already attached to the brief.
The illustrator is no longer necessarily being hired to invent the image. They may be being hired to make the machine’s image work.
That is a very different relationship between artist and client.
“Lazy and cheap”
Fashion illustrator Hayden Williams has been particularly blunt about this shift, describing the use of AI-generated imagery as “lazy and cheap.”
The phrase captures something at the heart of the backlash.
For many illustrators, the problem is not simply that AI can produce an attractive image. It is that clients can use the technology to bypass the creative labour that previously justified the commission.
Williams has also spoken about the importance of continuing to put human-made work into the world despite the rise of AI. In a recent interview, he said he was not going to let AI stop him from sharing his work, while arguing that there are still qualities in his own art that AI cannot reproduce.
His position reflects a wider frustration among illustrators: if the point of hiring an artist is creativity, judgement and individual expression, what happens when the client decides that generating hundreds of images with a prompt is good enough?
“Lazy and cheap” is therefore more than an insult directed at a technology. It describes the fear that AI could turn illustration into a commodity, where speed and cost matter more than the person who made the work.


Musicians face a different kind of AI threat
This does not mean musicians are less concerned about AI.
They have enormous reasons to be worried. AI can imitate voices, generate songs, reproduce musical characteristics and potentially compete with human musicians for commercial work.
A 2025 UK Music survey found that 66% of responding creators believed AI posed a threat to their careers.
And a broader 2026 report involving the Association of Illustrators, the Independent Society of Musicians, the Society of Authors and other creative organisations found that 73% of surveyed musicians said unregulated generative AI threatened their ability to earn a living.
So it would be misleading to say musicians are simply unconcerned about AI.
The difference may instead be how directly the technology attacks the individual creative product.
For a musician, AI can imitate a voice or generate music. But music already exists within an enormous technological ecosystem. Musicians have spent decades dealing with sampling, digital production, autotune, synthesizers, streaming and other technologies that alter what it means to make and distribute music.
Illustrators are confronting something slightly different.
Their finished product is already the thing the AI model is designed to generate.
Illustration has an especially uncomfortable relationship with style
There is also the question of style.
An illustrator’s style is not just a technical trick. It is often the result of years of experimentation, influences, professional experience and deliberate artistic choices.
Yet generative image models have made it possible for clients to ask for images resembling particular visual traditions or, controversially, particular artists.
That creates a peculiar economic problem.
An illustrator might spend years becoming known for a distinctive visual language. Suddenly, a client can type a description into an image generator and obtain hundreds of variations in seconds.
Even if the AI-generated images are technically inferior, they can still change what a client thinks an illustration should cost.
This is arguably one of the most psychologically damaging aspects of generative AI.
The issue isn’t only that a machine can make an image.
It is that the existence of the machine can cause the perceived value of making an image by hand to fall.
One illustrator quoted in reporting on the AOI findings described this as a “psychological devaluation” of the image.
That may explain some of the particularly strong hostility toward image-generating AI.
The problem goes beyond lost commissions
The most interesting part of the AOI survey may actually be what happens when illustrators do get the job.
If AI imagery is increasingly present in briefs and moodboards, the illustrator can end up working within a visual direction that has already been generated by a machine.
That potentially changes the creative process itself.
Instead of:
client → illustrator → finished image
the process can become:
client → AI-generated concept → illustrator → revised image
The illustrator is still employed, but their role has changed.
They may have less control over the concept, less time to develop it and less negotiating power over the fee.
This is important because creative jobs do not have to disappear completely for a profession to be damaged.
If the industry retains the final illustration but eliminates much of the conceptual, exploratory and developmental work that used to happen before it, the economic structure of the profession can still deteriorate.
And that matters particularly for people trying to enter the industry.
Junior illustrators traditionally build careers through smaller commissions. Those jobs may not be glamorous, but they provide experience, contacts, portfolio material and income.
If those entry-level jobs are increasingly replaced by cheap AI-generated material, the problem isn’t simply that fewer people get paid today.
It becomes harder for the next generation to become professional illustrators at all.
But are illustrators really more anti-AI than musicians?
The evidence does not allow us to make that claim definitively.
The AOI survey measures the experiences and opinions of its respondents, while music surveys use different samples and methodologies. Comparing percentages from different surveys as though they were one controlled experiment would be misleading.
In fact, musicians have demonstrated extremely high levels of concern about AI. The 2026 “Brave New World?” report found that 73% of surveyed musicians said unregulated generative AI threatened their ability to earn a living.
There is also evidence that musicians are not uniformly opposed to AI. A separate 2026 survey of 1,525 musicians found that 67% had used AI for music-related work, with professional musicians actually more likely to use it than hobbyists. At the same time, 55% of AI-using musicians in that survey expressed concerns about copyright and licensing.
That is an important distinction.
Being angry about how AI companies train their models is not necessarily the same thing as rejecting AI as a tool.
An illustrator might use AI for brainstorming while simultaneously opposing the use of their artwork in training datasets. A musician might use AI for stem separation while opposing AI-generated replicas of their voice.
The debate is therefore not simply artists versus AI.
It is increasingly about which forms of AI use artists consider legitimate.
Why the backlash feels so personal
Perhaps the biggest difference is psychological.
For many illustrators, generative AI can feel like a machine trained on the very thing they spent their lives learning to do.
The AOI survey found extremely strong opposition to the use of protected illustration for AI training. According to reporting on the survey, 98% of respondents wanted AI companies to seek permission before training on protected illustration, while 97% wanted transparency about what material had been used.
That makes the controversy fundamentally about more than automation.
It is about consent.
An illustrator can reasonably look at an AI-generated image and ask: Where did this visual knowledge come from? Who created the work that taught the system what illustration looks like? Were those artists asked? Were they paid?
Those questions are difficult to separate from the economic consequences now being reported.
The real divide may not be artists versus musicians
Ultimately, the distinction may be less about which creative profession “hates AI more.”
It may be about where AI enters the creative pipeline.
For illustrators, it can arrive at the beginning of the commission, inside the brief, inside the moodboard and eventually in place of the commission itself.
For musicians, the most explosive issues may instead involve voice cloning, ownership, streaming economics, rights management and the creation of synthetic performers.
Both groups are confronting the same underlying question:
What happens when technology can reproduce part of the thing you are paid to make?
The AOI’s latest figures suggest that, for illustrators, this question has already moved beyond speculation. Nearly half of the responding illustrators said they knew they had lost work to generative AI, while many others reported that AI was changing the briefs, budgets and deadlines attached to the work they still receive.
That may be why the hostility toward image-generating AI can feel particularly intense.
For an illustrator, AI isn’t merely a new piece of software sitting on the desk.
It can look like the client is sitting across the desk with a replacement already in their hands.
Illustrations by Hayden Williams
