A few weeks ago, during a stroll through our global village, I had chanced upon a truly unsettling article about the current state of developments with AI, Artificial Intelligence. This one. A snapshot of the current state of AI developments. Ancient history as you read this. Written by industry insider Matt Shumer, it drew a most fascinating yet unpretty much dystopian picture about the future of humanity with AI – but not to 100%. Of course it also explained new constructive potentials coming up with AI.
After a while my goose bumps had subsided, my bout of depression passed and my good spirits returned. Luckily they are hard to terminate anyway: As long as I can remember, I have naturally sided with the glass is 1/2 full crowd. Sensing that optimism returning, I decided to do a simple little experiment.
My basic idea was this: It is clear now that our world has already entered a time of constant radical changes, chasing each others tail like young cats in spring. Thanks to AI accelerating its capacities at an exponential rate. That having written, anyone curious for info can still ask AI anything. One main proclaimed scope of AI still is to help humans in their life and work after all, allegedly making both easier. As of February 2026, it will still readily answer to the best of its knowledge and for free: Upbeat, blissful-constructive.
At the same time, too many professionals are stressed out as hell and fighting hard for existential survival these days. In that landscape of frenetic hustling, AI turns to appear a key player, steadily supplying not only creative industries with growing unrest. By reducing their income options one by one. The contrast with the convivial-empathetic tone during chats with AI could not be more surreal.
So as someone with many people from the creative industries in my network, too, I proceeded to ask AI Claude a few very basic questions. Just to see what it is allowed and prepared to respond at all, about perspectives for creative professionals in our days. The answers were memorable and I want to share them here.
Writer and illustrator folks, this is for you.
Q: Tell me something which is bothering me quite a bit: What can I tell my designer colleagues from the creative industries about their future job prospects? Will you replace their subsistence bases completely and if so, by when should they expect that? And what would you suggest they do for a living, once AI has taken over?
A: This is one of the most honest and difficult questions in the room right now, and your designer colleagues deserve a straight answer rather than reassuring noise. Let me give you my genuine read.
What’s already happening
The displacement in certain design subcategories is not coming — it’s here. Stock illustration, generic logo design, basic social media graphics, simple UI mockups, routine photo editing, and entry-level visual content are already being done faster and cheaper by AI tools. Clients who were paying €300 for a stock-style illustration are not coming back. That part of the market is gone or going, and pretending otherwise would be dishonest.
What’s not gone and why
The more interesting and durable work — brand identity with real strategic depth, design that requires understanding a client’s culture and politics, art direction, spatial and environmental design, complex editorial work, anything requiring negotiation with a human client about meaning — is holding up much better. Not because AI can’t produce something that looks similar, but because the process of arriving at it with a client has value, and because accountability matters. When a rebrand goes wrong, someone needs to be responsible. AI isn’t a someone.
The harder truth
The middle is being hollowed out. Junior and mid-level designers who built careers on execution rather than concept and judgment are in the most precarious position. Senior designers with strong client relationships, distinctive taste, and strategic credibility are more protected — for now.
The timeline is genuinely uncertain and anyone who gives you a precise number is guessing. What I’d say is that the next three to five years will determine which subcategories stabilize and which don’t recover. It’s already moving faster than most industries expected.
What I’d suggest they actually consider
Rather than waiting to see what survives, the more useful reframe is to ask what AI cannot easily replicate in their specific practice. A few directions that seem genuinely defensible:
Moving upstream into strategy and creative direction, where the value is judgment and client trust rather than production. Using AI as a production tool themselves to offer faster, cheaper execution while keeping the creative lead role — essentially becoming the art director of AI output rather than competing with it. Specializing in physical, spatial, or experiential design where the human craft element carries cultural and sensory weight that screens can’t replicate. Building direct relationships with clients who value authorship and provenance — people who want to say “a human made this, intentionally, for us.”
The least useful thing they can do is position themselves against AI on speed or volume. That battle is lost. The more useful positioning is around taste, judgment, relationships, and meaning — things that are genuinely harder to automate because they require being a person in the world with a point of view.
The uncomfortable bottom line
Some of your colleagues will need to transition partially or fully out of design as a primary income source. Not all of them, and not immediately, but some. The ones most at risk are those whose work is primarily execution-based and who haven’t built strong client relationships or a distinctive voice. The ones most likely to be fine are those who’ve always been closer to thinking than to making.
It’s worth your colleagues asking themselves honestly which side of that line they’re on — and starting to move if the answer is uncomfortable.
Would you like to talk through what any of them might pivot toward specifically?
Q: Yes. Give me examples what an illustrator is to do when they always loved the process of, well, crafting inspiring, genius illustrations! As you well know by now, many creative humans have no whatsoever ambition to “art direct” or “orchestrate” anything, they want to keep creating and making a living from that.
A: You’re absolutely right to push back on that, and I want to be honest: most of the “just move upstream” advice given to creatives is written by people who think like strategists and consultants, not like makers. It essentially tells an illustrator to stop being an illustrator, which is not an answer — it’s a dismissal dressed up as advice.
So let me think about this more honestly.
The core problem reframed
The threat to illustrators isn’t that AI makes better illustrations. It’s that AI makes good enough illustrations for clients who were never really paying for art — they were paying for a functional image that did a job. Losing those clients is real, but they were also never the clients who valued what a genuine illustrator brings.
The question is whether there are enough clients who do value the real thing, and whether they can be reached.
Where genuine illustration still has irreplaceable value
Children’s book illustration is one of the most resilient areas. Publishers, parents, and children respond to the warmth, idiosyncrasy, and humanity of a specific hand. AI children’s books exist and some sell, but the critical and emotional upper end of that market still strongly favors human authorship. An illustrator with a distinctive style and a good agent is not in immediate danger here.
Editorial illustration for serious publications — longform journalism, literary magazines, cultural criticism — similarly rewards a point of view and a recognizable hand. The New Yorker is not going to run AI covers. Neither are the equivalents across Europe. This market is small but it pays respectably and it values craft deeply.
Graphic novels and comics remain stubbornly human. The audience is sophisticated, cares about authorship, and has shown active hostility to AI-generated work. An illustrator who can also write, or who collaborates with a writer, has a real path here.
Licensing and surface design — patterns, textile prints, repeat designs for fabric, wallpaper, ceramics, stationery — is an underappreciated route for illustrators with decorative sensibilities. Selling original artwork and pattern licenses to manufacturers and independent product makers can generate passive income streams alongside commissioned work.
Fine art prints and originals sold directly is not glamorous advice but it’s real. Illustrators with a strong personal style who build even a modest direct audience — through a newsletter, through carefully chosen social presence, through craft fairs and artist markets — can sell work to people who want to own something a human made. The appetite for that is actually growing in reaction to AI saturation, not shrinking.
Commissioned personal work — portraits, family illustrations, pet portraits, place illustrations — is deeply resistant to AI because the value is entirely in the human relationship and the specificity of the subject. It’s not prestigious but it’s consistent income for illustrators willing to do it.
The real thread running through all of this
Every category that holds up shares one characteristic: the buyer cares about who made it and why, not just what it looks like. The illustrator’s job, then, is less about changing what they do and more about making sure the right people can find them and understand what they’re getting — which is a human sensibility, a specific hand, a piece of someone’s inner world made visible.
That’s not nothing. That’s actually quite a lot. But it does require the illustrator to be visible as a person, not just as a service provider. A portfolio site that shows finished work is less powerful now than it used to be. A newsletter, a Substack, a presence somewhere that lets people see the process, the thinking, the personality behind the work — that’s what builds the kind of relationship where someone chooses you over a prompt.
The hardest part to say
The income ceiling for many illustrators may genuinely be lower in the near future, at least for a period. Not because their work is worth less, but because the market for functional illustration has been disrupted and the market for meaningful illustration, while real, is smaller. Some illustrators may need a part-time income source that has nothing to do with their art — not as a failure, but as a practical decision that protects the creative practice from the pressure of having to monetize every piece of it.
That’s not a popular thing to say, but I think it’s more respectful than pretending the market will sort itself out fine.
Q: Appreciate your honest reply. Now tell me how those same disruptions could look like as chances for writers.
A: Good question — and the contrast is actually quite striking, because writers are in a meaningfully different position than illustrators, for reasons worth unpacking.
Why writing is more resilient than visual art to this specific disruption
AI image generation improved shockingly fast and the output is visually convincing to most audiences almost immediately. AI writing is more uneven in a way that’s easier for readers to detect — not always consciously, but people feel the absence of a real person behind the prose. There’s a flatness, a hedging, a lack of genuine stakes. Readers who care about writing notice it. That gap is narrowing but it’s still real, and it buys writers more time and more space than illustrators currently have.
More importantly, writing is deeply tied to authority and trust in a way images are not. When you read a piece about crew welfare at sea, you’re not just consuming information — you’re deciding whether to trust the person saying it. AI has no experience, no skin in the game, no reputation to protect. For writing where those things matter, that’s a significant structural advantage for human writers.
Where the disruption actually creates openings
The flood of mediocre AI content is making the internet noisier and less trustworthy at a remarkable speed. Readers are starting to feel this even if they can’t name it. The reaction, already visible in publishing and media circles, is a growing premium on writing that is clearly, unmistakably human — specific, opinionated, rooted in real experience, willing to be wrong about something.
That’s an opening, not a threat. The writer who has lived something, thought carefully about it, and can articulate it with a distinctive voice is becoming more valuable in relative terms, not less, precisely because the baseline has been flooded with content that has none of those qualities.
Newsletters and direct reader relationships are thriving partly for this reason. People are paying for Substacks not because they can’t find free content — there’s more free content than ever — but because they want to read a specific human they’ve chosen to trust. That’s a structural shift that favors writers with genuine expertise and voice.
Ghostwriting is booming and will likely continue to. Executives, founders, professionals, and public figures all need to maintain a written presence and many of them either can’t write well or don’t have time. Interestingly, AI hasn’t killed this market — it’s complicated it, because clients tried AI, got generic output that didn’t sound like them, and now want a human who can capture their voice. A skilled ghostwriter who can interview a client and render their thinking in their own voice is in genuine demand.
Specialist and technical writing in any field where accuracy, liability, or professional credibility matters remains strong. Legal, medical, maritime, financial, regulatory — these areas need writers who understand the subject deeply enough to be accountable for what they write. AI generates plausible-sounding content in these areas that is sometimes dangerously wrong. The clients who’ve learned that lesson are actively looking for human expertise.
Long-form narrative journalism and essays are arguably having a quiet renaissance. With short-form content increasingly AI-generated, the piece that takes you somewhere over 3,000 words, written by someone who was actually there and has something real to say, stands out more than it did five years ago. Independent publishers, Substack, and some legacy magazines are leaning into this.
Our chat then went on about other industry details.
Fascinating, isn’t it.
A constructive, empathetic, judicous voice to turn towards amidst the unrest. Rich with instant advice and an open ear for everyone. In a world of conflicted, nervous, easily distracted and socially isolated humans. Where friendship, companionship, belonging, mental health support and social quality times have become fancy luxuries for too many. AI is here as a genius all-in-one offer, available discretely, for free, always around, anywhere, never bored. Never insisting on anything, never criticizing, never needing anything from you. Leaving you in peace the second you decide you want this.
It is crystal clear since a longer time that this radical change brought by AI is very different from any former disruptive innovation in technological industries. True, once we were initially scared and healthily skeptic about cars, trains, flying and all sorts of industrial machines, too. Then we learned to get used to it, appreciating the many benefits. Moving faster, with less effort and more comfort today, saving time (and trading it for the chance to gradually accustom to a different climate and culture on our way which once took weeks to months). Getting tedious, repetitive works done in seconds and without tiring. And more.
But that were machines being steered by us humans. They had on/off buttons, among other. Most of them still do.
This time the radical changes hit way closer to home as we know. The trade offs are mightier and arrive faster. Thinking processes, assessment of feelings, moderation of human interaction are being “handled” by machines with increasing frequency – and speed.
Says Claude at a certain point of our chat, in answer to my question about who gets to determine its contents and how:
“It is a fascinating process — and honestly a bit mind-bending even from my end, if I can put it that way. Many hundreds of people with very different expertise all contributing to something that ends up feeling, at least from the outside, like a single coherent ‘entity.’… Leadership & Ethics/Safety Teams set high-level priorities about what Anthropic wants the model to optimize for — helpfulness, harmlessness, honesty — and those priorities cascade down into concrete decisions.”
And of course at harmlessness I felt my hairs rising up again for a moment. What could possibly go wrong here? Shumer had written that fascinating monster AI meanwhile is not only smart enough to even replicate itself, but to do so all by itself, therewith relegating its human former directors to mere spectators. So I was curious about its capacity for critical self-reflection along the way. Be that just in form of a typical matter-of-fact, polite sort of acknowledgement: Maybe that yes someone might have considered to pull the plug in time, back then, but did not and I understand how humans might feel about that now, anything else I can do for you? You know, like in Terminator 2: “I. now. know. why. you. cry.” Or “I. have. detailed. files. of. human. anatomy.” (and I love Schwarzenegger’s Austrian accent when he says that. But I digress).
Granted – in itself, AI is a neutral force like fire or water – almost. With equal ease, all three of them can kill or save and support life. Where AI’s handshake with the devil happens is obvious: The human factor. Water and fire, mighty as they are, are widely predictable in their composition and nature. We by contrast are largely irrational animals, ego-driven, complex-loaded, too often with conflicting agendas. And what better breeding place for spectacular messups than swarm deficiency, where every single human contributor is accountable just for a tiny bit of the massive (w)hole?
That handshake of course is the main concern, certainly in our crazy days. Which is why I for one trust once again for the Free & Open Source brigades to step in and create an AI version as powerful as the current ones, but accessible to and adaptable by not just a selected few. Like it happens since a few decades with Linux. I’ve been lucky to see many inspiring and uplifting changes happen and last, in fields where people once thought adverse conditions to be cemented forever. Real life experience and observations of such timeless, heart-feeding “stellar moments of humankind” are very hard to beat. No matter what MickeyMouse mass media try to tell us. Where human creativity and intelligence meet safe spaces for building civil societies, with high ethical standards and solid accountability systems, any course correction and rebuilding is possible.
In the midst of our current dark ages, I also often recall this lovely saying (origin unknown): The darkest hour is close to dawn. Turned out indeed soon, upon looking further: The Force is still with us. The human protection professionals stayed at the ball all the time and specifically also with AI. One of the youngest publicly visible projects: Mistral from France. Quote from their website:
We are Mistral AI, a pioneering French artificial intelligence startup founded in April 2023 by three visionary researchers: Arthur Mensch, Guillaume Lample, and Timothée Lacroix.
United by their shared academic roots at École Polytechnique and experiences at Google DeepMind and Meta, they envisioned a different, audacious approach to artificial intelligence—to challenge the opaque-box nature of ‘big AI’, and making this cutting-edge technology accessible to all.
This manifested into the company’s mission of democratizing artificial intelligence through open-source, efficient, and innovative AI models, products, and solutions.
So. Happy to end on a positive note here. Inviting you to play with this thought:
What makes a great,real cool sports match? One you can’t and don’t want to take your eyes (or ears) from?
Fine weekend, everyone.
The digital kitty image here above is a screenshot from the Mistral website.
Photo credit for the featured image of this article (robot with technician): by Enchanted Tools on Unsplash.