The Market for Imperfection
Visible brushstrokes command a premium in the age of synthetic imagery.
The conventional wisdom has been clear enough: generative AI is standing over commercial illustration with a knife. The logic seemed airtight. Why pay an illustrator when Midjourney can produce serviceable images in seconds? Why commission original work when stock libraries overflow with AI-generated alternatives at negligible cost?
And yet, talking to illustration agents and artists at the close of 2025, a different picture emerges. Some illustrators and their agencies actually thrived last year. Not despite the AI onslaught, but in a certain sense because of it.
The field is not dying, but it is changing rapidly and in directions that confound the simple displacement narrative. What one hears from industry insiders is less a funeral dirge than a call for reinvention, an insistence that the moment demands something human-made imagery can provide that AI cannot: intention, idiosyncrasy, and unmistakable human presence.
The clearest trend is the resurgence of handmade techniques: woodblock printing, linocuts, hand-lettering, visible brushstrokes, paper textures, uneven linework. The Central Illustration Agency in London has just signed a Japanese woodblock printmaker to its roster. IllustrationX recently added linocut artist Emily Robertson. These are not nostalgic gestures but competitive responses to a market saturated with algorithmically generated smoothness.
The reason is straightforward: AI image generators excel at producing polished, technically competent imagery but are less capable of producing work that carries what industry insiders call the “hand of the artist,” the visible evidence that a specific human being made choices, left marks, and invested time. As Mohamed Danawi, founder of the agency Illozoo, told Creative Boom: clients and consumers are increasingly looking for work with “intention, emotional resonance and human insight.” They want imagery that “feels authored, thoughtful and genuinely communicative.”
This represents a meaningful shift in what illustration is for. For decades, much commercial illustration served a utilitarian function: fill this space, visualize this concept, decorate this package. AI can handle utilitarian illustration competently enough. What it cannot do is bring a distinctive point of view, cultural specificity, or the kind of personality that makes viewers stop and actually look.
Kate Sherwin at the design firm FutureBrand put it plainly in a recent interview: in 2026, the big trend will be to “embrace the mess.” The phrase captures something essential about the moment. People are tired of pixel-perfect imagery, of the frictionless surfaces that AI produces so easily. They want work that shows evidence of process, of human decision-making, of craft.
This is visible across the industry: Spotify’s 2025 Wrapped campaign featured hand-drawn illustration, Oatly continues to build its brand identity around scribbly type and intentionally uneven linework, and the British pharmacy chain Boots recently partnered with ARTHOUSE Unlimited, highlighting work by artists with diverse learning and physical disabilities whose value lies precisely in its deviation from algorithmic norms.
Jamiel Law, a senior designer at the creative studio BUCK, observes that audiences are “hungry for visuals that feel distinct, human, and authentic” precisely because flat, corporate illustration has reached saturation. The sheer volume of AI-generated imagery has created a countervailing demand for something it cannot provide.
Artists themselves are responding by diversifying. Reports from trade publications describe digital artists learning 3D software, exploring AR and VR, turning to game design, or pivoting back to traditional media like painting and ceramics. Video game artist Michal Gutowski, to take one example, now also runs a pottery studio. These are not retreats from the industry but expansions, attempts to build multiple income streams and to occupy creative territory that remains resistant to automation.
The pivot to traditional media is particularly striking. For years, the trajectory in commercial art moved consistently from analog to digital, from physical materials to software. Now that movement is reversing in certain pockets of the field. Some of this reflects practical calculation, since original paintings and prints can be sold as objects rather than merely licensed as images, but some of it reflects a deeper recalibration of what makes creative work valuable in an era of infinite synthetic reproduction.
The illustration agent Marie-Christine Brisson, co-founder of Colagene Creative Clinic, captures the mood: “One thing is certain, a return to handcrafted styles will come.” But she adds something more revealing: “And I hope to see many new forms of expression. We all need to be surprised again!”
That desire for surprise, for genuine novelty, points toward what AI cannot easily deliver. Generative models produce variations on existing patterns. They optimize for statistical likelihood, which tends toward the familiar. A recent study from Dalarna University demonstrated this empirically: when researchers let AI image and text systems iterate autonomously, the outputs rapidly converged on a narrow set of generic visual styles, regardless of starting point. The researchers called the result “visual elevator music,” pleasant and polished but devoid of meaning.
Human artists, by contrast, can make strange choices, combining influences that have no statistical relationship to each other and pursuing aesthetic intuitions that emerge from specific life experiences. This capacity for the unexpected is increasingly what clients are paying for.
None of this means the industry is healthy or that displacement is not occurring, because it clearly is. Making a living as an illustrator grew harder in 2025, and the pressure from AI-generated imagery is real, particularly at the commodity end of the market where generic images serve generic purposes.
The accurate picture is probably one of bifurcation: commoditized illustration, the stock-image-style work that fills websites and advertisements without much thought, will increasingly be automated, while bespoke, personality-driven work that depends on a specific artist’s vision will retain premium pricing for those who can demonstrate their value.
The middle of the market is where the squeeze happens, as illustrators who produced competent but undistinctive work, who filled briefs adequately but without memorable style, face the most pressure because they are competing directly with tools that can produce adequate work for almost nothing.
What seems to be emerging is a clearer separation between illustration as commodity and illustration as authorship, where the former is increasingly automated while the latter is increasingly valued precisely because it is not.
For artists navigating this landscape, the advice from agencies is consistent: develop a distinctive voice, cultivate cultural specificity, and demonstrate unmistakable human presence in the work itself. Ana Bandarra, the IllustrationX representative in Brazil, says clients now want to commission artists to collaborate on projects “with narratives that come from where they stand, adding their authentic points of view, from a place of belonging in the specific universe that clients are communicating within.”
This is a significant change from the era when illustration was primarily about execution. Now the person behind the work matters as much as the work itself. Lived experience, cultural background, individual perspective become part of the value proposition.
Whether this represents a sustainable path for the industry or merely a temporary respite remains uncertain, as AI models continue to improve and the definition of what counts as distinctively human may shift as synthetic imagery grows more sophisticated.
But for now, the human hand is fighting back, and it is fighting not by competing with AI on AI’s terms, not by producing more images faster, but by emphasizing what remains difficult to automate: the peculiar, the personal, the unmistakably made.
The mess, in other words, and the imperfection, and the evidence that someone was here.


One of the biggest advantages of AI is not having to deal with artists. They are some of the most short-sighted, unreliable, and dishonest people on the planet.
Forget quality. Just getting the work one paid for delivered doesn't happen at least 15 percent of the time. Arkhaven learned the hard way why Marvel and DC will pay writers advances, but never illustrators.
So... how long until we can program AI to imitate the "hand of the artist"?