AI and Creativity: How AI Is Commoditising Average Creative Output

For years, the conversation around AI and creativity has been forced into a rather lazy binary. On one side, the optimists promised a world of unlimited creative possibility, where anyone could become an art director, copywriter, designer, filmmaker, strategist or composer with the right prompt. On the other, the pessimists warned of a soulless machine age in which human imagination would be scraped, simulated and eventually made redundant.

Both sides had a point. Both sides also missed the more immediate and uncomfortable truth.

AI is not killing creativity. Not really. Not yet. What it is doing, with frightening speed and rather little sentimentality, is commoditising average creative output.

That distinction matters because “average” has quietly carried a very large part of the creative economy for a very long time. Good enough copy. Good enough layouts. Good enough campaign concepts. Good enough social posts. Good enough moodboards. Good enough brand worlds. Good enough content calendars. Good enough assets that kept clients supplied, agencies billing, freelancers working and in-house teams looking efficient.

The problem, of course, is that “good enough” is exactly where AI is strongest.

Stanford’s 2025 AI Index found that 78% of organisations reported using AI in 2024, up from 55% the previous year, while 71% reported using generative AI in at least one business function. In other words, this is no longer a speculative debate about whether the tools will enter professional workflows. They already have.

The question now is not whether AI can produce creative work. It can. The sharper question is what happens to creative value when the baseline becomes cheap, fast and abundant.

Peter Stylianou, Creative Lead Director at PeteSake Designs, puts it about as directly as anyone could:

“AI is not killing creativity. It is killing average creative output. For years, a lot of creative work survived because it was ‘good enough.’ Good enough copy, good enough layouts, good enough social content, good enough campaign ideas. Now AI can produce ‘good enough’ in seconds.”

That is the thesis. Not the end of creativity, but the end of “good enough” as a defensible business model.

Why AI and Creativity Are No Longer Opposites

Suhail Chaudhary

The old assumption was that AI and creativity belonged in separate rooms. AI was logic, automation and efficiency. Creativity was taste, intuition and emotional intelligence. AI was the spreadsheet; creativity was the spark.

That distinction now feels increasingly antique.

AI has become part of the creative process because much of the creative process, particularly at the commercial end, involves tasks that can be described, templated, varied and iterated. Drafting. Resizing. Referencing. Exploring territories. Producing alternates. Testing headlines. Generating visual directions. Summarising research. Creating first-pass scripts. Building presentation structures. Producing thumbnails. Reworking tone. Extending campaigns across formats.

None of that means AI has suddenly developed taste, memory, cultural instinct or a meaningful relationship with the messiness of human experience. It means it can generate plausible work at scale. And in a commercial environment that often rewards speed and volume before depth and distinctiveness, plausible is enough to cause disruption.

The UK government’s own sectoral work on AI skills has already identified generative AI use across the creative industries, including content creation, campaign planning and digital storytelling, while also noting the need for creative professionals to evaluate AI outputs for originality, quality and audience alignment.

That last part is crucial. The future creative skill is not simply the ability to operate the machine. It is the ability to judge what the machine gives back.

Stylianou gets to the heart of this when he asks whether some of the work now being replicated so easily was ever valuable in the way the industry liked to imagine:

“That is uncomfortable, but it also forces the creative industry to ask a necessary question: If AI can replicate the work that easily, was the value really in the output itself?”

It’s a brutal question, but a useful one. Because for all the romantic language the industry uses about ideas, craft and originality, quite a lot of its commercial machinery has depended on repeatable output. A client needs twenty captions. A brand needs ten visual routes. A marketing team needs a month of LinkedIn posts. A pitch deck needs polishing. A retailer needs endless product copy. A junior designer needs to turn one campaign into fifty formats by Tuesday.

That work still matters. It still takes judgement to do well. But it is also the kind of work AI can now attack first because it sits close to pattern recognition.

As Stylianou says:

“The real shift happening now is that basic execution is becoming cheaper, faster, and more accessible. AI can generate decent options, but it still does not truly understand taste, brand memory, emotional context, cultural timing, or strategic relevance. That is where human creativity still matters.”

This is why AI and creativity are no longer opposites. AI has entered the room not as a replacement for all creative intelligence, but as a force that changes which parts of creative labour are scarce. Execution is becoming less scarce. Judgement is becoming more important. Output is becoming abundant. Taste is becoming the differentiator.

How AI Is Raising the Baseline for Creative Output

Lark

Is AI making creativity less valuable, or simply more accessible?

The honest answer is both.

It is making baseline creative production more accessible to people who previously lacked the technical skills, time or confidence to produce polished work. That is not nothing. There is a democratic argument here. Small businesses, charities, junior marketers, solo founders and non-specialist teams can now make things that look substantially better than they might have managed before.

But accessibility has a market consequence. When more people can produce competent work, competence itself becomes less valuable.

A 2024 study from the University of Exeter and UCL found that AI assistance improved the perceived creativity and quality of short stories, with the strongest gains among less creative writers. The same study also warned that AI-assisted work may become less varied overall, because generated assistance can pull people towards similar patterns.

That is the central paradox. AI can make individual pieces of work look better while making the wider creative landscape feel flatter.

Simon Manchipp, Founder at SomeOne, phrases this with characteristic economy:

“Average used to be a business model. Now it’s a button.”

That line should probably be printed out and pinned above a few agency finance departments. Because it identifies the real commercial shock. The machine does not need to beat the best creative talent in the world to change the market. It only needs to beat the economic case for average work.

Manchipp continues:

“If your creative work is okay, congratulations: you are now competing with a machine that does okay for free, in three seconds. If you aren't aiming for exceptional, you might as well pack up. Creativity is now on the borders of being a luxury purchase.”

There is a discomfort in that phrase, “luxury purchase,” because creativity has long liked to think of itself as essential. And at its best, it is. Distinctive brand thinking, culturally fluent campaigns, beautifully crafted design systems, sharp strategy and original storytelling are not decorative indulgences. They are forms of commercial advantage.

But average creativity? The sort that fills the calendar, satisfies the brief, avoids risk and produces something familiar enough to pass? That is exactly the kind of thing clients may increasingly see as optional human labour.

Shagorika Heryani, Founder at Athina, pushes this even further:

“Average is now free and mediocre by design. AI was trained on the accumulated average of human history and is structurally built to give you the most statistically likely answer. The most pleasing colour palette. The safest content. Accurate, yes. Efficient, yes. But average is its ceiling, not its floor.”

That is a vital point. AI does not merely produce average work by accident. In many cases, average is the logic of the system. Generative AI is built to predict, assemble and produce outputs that fit recognisable patterns. It is brilliant at the plausible middle. It understands what usually comes next. It knows what a brand manifesto tends to sound like, what a fintech landing page tends to look like, what a lifestyle ad tends to say, what a pitch headline tends to do.

The danger is that this plausibility can be mistaken for quality.

Heryani’s provocation is not that AI is useless. Far from it. It is that AI’s greatest strength is also its creative limitation:

“because if average is a commodity, then genuinely original thinking just became both harder to produce and far more valuable.”

That is where the baseline shift becomes interesting. The average rises. The middle gets crowded. The surface improves. But genuine originality becomes more exposed, more necessary and more difficult to fake.

Why Average Creative Work Is Becoming a Commodity

Larasita Mahardhika

To say average creative work is becoming a commodity is not to insult the people who do it. It is to describe a market shift.

A commodity is something broadly interchangeable. One supplier’s version may be slightly better or worse, but the buyer does not perceive enough difference to justify a significant premium. That is what threatens the middle of the creative market. Not because clients hate creativity, but because AI gives them access to outputs that look close enough to what they thought they were buying.

Becky McOwen-Banks, Founder at Plain:AI, is especially sharp on this point:

“Let’s start with the uncomfortable truth that most of the industry is still dancing around. AI isn’t coming for the best creative work. It’s already eaten the middle.”

That middle is where the impact is likely to be most profound. Not the world-class identity system. Not the emotionally devastating film. Not the campaign that changes public behaviour. Not the illustrator with a truly singular hand. Not the strategist who can reframe a business problem in a way that unlocks growth.

The middle.

As McOwen-Banks explains:

“For decades, the creative industry ran on a pyramid. At the top, genuinely original, distinctive, culture-shifting work. At the bottom, templated, production-line output. And in the middle — a vast, profitable layer of competent, decent, good-enough creative work that kept agencies staffed, freelancers booked and careers ticking along nicely. That middle is collapsing. Not gradually. Quickly.”

This is the part of the conversation that often gets softened because nobody wants to sound cruel. But it is better to be blunt. A lot of creative businesses were built on work that was never exceptional but was professionally competent, reliably delivered and difficult enough to produce that clients paid for it.

Now AI can generate competent copy, workable concepts, adequate design routes and passable video treatments at a fraction of the cost and speed. As McOwen-Banks says:

“Not better than the best humans. But better than average — and faster and cheaper than anyone can compete with on pure economics.”

That phrase “on pure economics” is doing a lot of work. It explains why this is not simply a craft debate. It is a procurement debate. A staffing debate. A margin debate. A freelance-rate debate. A training-pipeline debate. A question of how agencies justify their fees when clients can see an approximation of the thing appear almost instantly.

It also explains why so many clients are tempted to bring production in-house or run AI tools directly. The risk, of course, is that they may know how to generate output without knowing how to recognise value.

McOwen-Banks captures that danger neatly:

“When a client runs an AI tool without genuine creative expertise in the room, they get output that is technically functional and creatively hollow. It meets the brief at surface level. It ticks the boxes. It looks like the thing without being the thing. And here's the problem — they often can't tell.”

This is where commoditisation becomes insidious. Bad work is easy to reject. Hollow work is harder because it often looks fine. It has the right structure. The right vocabulary. The right colours. The right rhythm. It resembles the category. It resembles previous work. It resembles competence.

But resemblance is not the same as resonance.

The output may land flat. The brand may blur. The campaign may disappear into the feed. The illustration may feel strangely lifeless. The copy may be polished but personality-free. The concept may be familiar enough to pass a meeting but not strong enough to survive culture.

McOwen-Banks again explains:

“Creative judgement — knowing what's missing, what's off, what needs to change and why — is not something you can replicate by giving a non-creative a good prompting guide. It's built through years of making things, failing, refining, developing an eye and an instinct that operates faster than conscious thought.”

That is the industry’s challenge now. Not to insist that AI cannot produce anything useful, because it plainly can. But to explain the value of judgement in a market dazzled by output.

What AI Means for the Creative Industries

Nalla Design

The phrase AI in the creative industries covers too many realities to be reduced to a single verdict. For some, AI is a productivity tool. For others, it is a threat to livelihoods. For some brands, it is a way to test, version and scale. For many creators, it is an extraction machine trained on work they never consented to provide. For junior creatives, it may be both a superpower and a trap. For agencies, it is both a margin opportunity and a direct challenge to the services they once sold.

The copyright battle is already fierce. In March 2026, the House of Lords Communications and Digital Committee warned that UK creative industries face a “clear and present danger” from generative AI, calling for stronger licensing, transparency and protection for creators. Reuters has also reported on the UK government’s attempt to reset the copyright debate after widespread backlash from creative stakeholders over earlier opt-out proposals.

That legal and political context matters because the creative industries are not merely reacting emotionally to new tools. They are asking a basic question: if AI systems are trained on human creative labour, and then used to compete with human creative labour, who benefits?

Andy Howell, Creative Director at The Clearing, brings the issue into focus through illustration:

“We’ve been having some tough conversations lately. The world-class illustrators we work with – people who can capture a brand’s entire ethos in a single stroke of a pen – are watching their commissions dry up. We’re seeing ‘tightening budgets’ used as an excuse to replace soul with software.”

That phrase, “replace soul with software,” could sound melodramatic if the stakes weren’t so real. Illustration is one of the most exposed areas because the output is highly visible, highly promptable and often misunderstood by clients as an image rather than as a point of view.

Howell explains why The Clearing has drawn a line:

“At The Clearing, we won’t be using AI for illustration. Here’s why we think the industry needs to wake up before we lose something we can’t get back. When we commission an illustrator, we aren’t just buying an image. We’re buying a point of view.”

That sentence is really the whole argument for human creative value. A point of view cannot be reduced to the final file. It includes lived experience, taste, restraint, references, decisions, accidents, obsessions, discomforts, cultural memory and the artist’s relationship with the brief.

AI can simulate styles. It can produce images that appear finished. But Howell is right to identify the difference between visual output and authored perspective:

“AI works on averages. It looks at everything that has already been done and gives you a ‘mathematical middle’. But great branding doesn’t live in the middle, it lives at the edges. It’s the intentional wobble in a line, the unpredictable choice of colour, the visual metaphor that makes you stop and actually feel something. That doesn’t come from data, it comes from hard earned experience.”

This is where the AI debate becomes a brand debate. If every brand uses the same tools, trained on overlapping data, prompted by people reading the same LinkedIn posts about prompt engineering, how long before everything starts to feel faintly related?

Howell warns of exactly that:

“If everyone uses the same AI models, every brand will start looking like a hallucinated fever dream of the same five ideas.”

John Keough, Senior Designer at Further New York, makes a similar point from the perspective of design language:

“Everyone can design something now. Logos, posters, campaigns, and brand identities are made by non-designers in an afternoon. Their outputs look real. Like something you've seen before. The tool they're using is built to replicate our current lexicon. Everything that's already been made. They're not creating a new brand. They're making a brand look like everyone else's.”

That is perhaps the most dangerous phrase in the whole discussion: “Their outputs look real.”

Because that is enough to fool a lot of people for a while. It is enough to get through a meeting. Enough to fill a slide. Enough to impress someone who does not know the difference between visual finish and strategic distinctiveness.

But brands are not built by looking vaguely like brands. They are built by becoming recognisable, memorable and meaningfully different.

Keough continues:

“AI's output is average because that's what acceptance looks like. Breaking that accepted standard takes mastery of the underlying system. If you want to move the lexicon forward, don't outsource the thinking. The tool can only ever give you where the industry already is, you'll always be one step behind. The tool knows every rule and can break none of them. Until you do.”

That is a beautiful way of putting it. AI can reflect the current lexicon. It can remix what already exists. It can help you understand the codes of a category. But moving the lexicon forward requires someone willing to understand the rules deeply enough to violate them with purpose.

The New Competitive Advantage is Originality

Digital Da Vincis

If AI makes production easier, then production alone becomes less defensible. This does not mean craft no longer matters. It means craft has to be attached to something more ownable: originality, strategy, taste, cultural awareness, emotional precision, brand memory and judgement.

The World Economic Forum’s Future of Jobs work has identified creative thinking, analytical thinking, technological literacy, resilience and lifelong learning as increasingly important skills. OECD work has also pointed to rising demand for originality-related skills in AI-exposed occupations. The direction of travel is clear enough: as tools absorb more routine execution, the human premium shifts towards higher-order judgement.

Stylianou states it simply:

“The future value of creative professionals will not simply be the ability to make things. It will be the ability to judge what is worth making. AI can generate 100 ideas. A strong creative knows which 99 to kill.”

That is one of the most useful definitions of creative value in the AI era. Not generation, but selection. Not volume, but discernment. Not the ability to produce options, but the ability to recognise the one worth pursuing.

This is where AI can actually make creative people more valuable, but only if they move away from defending the wrong thing. The industry cannot win by arguing that machines cannot produce outputs. They can. It cannot win by pretending every human-made piece of work is inherently better. It isn’t. It cannot win by treating AI fluency as beneath “real” creatives. That would be complacent and self-defeating.

The argument has to move up the chain.

McOwen-Banks is especially strong here:

“Commoditisation is not the end of creative value. It's the end of undifferentiated creative value. And that's actually an opportunity, if you're willing to move towards it rather than away from it.”

That is the key distinction. AI threatens undifferentiated creative labour. It does not eliminate the need for sharp creative intelligence. In fact, it may increase the premium on it.

McOwen-Banks continues:

“The work that AI cannot commoditise is work that is genuinely, specifically, irreducibly human. Work that carries a distinct point of view. Work that understands cultural nuance at a level a probability engine can't reach. Work that takes creative risk rather than optimising for the expected output.”

The phrase “optimising for the expected output” is useful because it describes so much AI-generated creative work. It tends to produce what feels right because it has learned what “right” usually looks like. But much of the best creative work does not feel right at first. It feels awkward, dangerous, funny, too simple, too weird, too quiet, too rude, too emotional, too specific, too early.

Originality is often uncomfortable before it is obvious.

That is why the best creatives do not merely answer briefs. They interrogate them. They notice the hidden problem. They reject the category cliché. They find the uncomfortable truth. They know when a brand should speak and when it should shut up. They know when polish is killing the idea. They know when the ugly route is more memorable than the tasteful one.

McOwen-Banks again continues:

“It also cannot replace the strategic thinking that shapes what gets made in the first place. The brief interrogation. The reframe. The question that changes the direction of an entire campaign. That's where experienced creative minds earn their place — not in the execution layer that AI is rapidly making its own.”

That is where agencies and freelancers need to reposition themselves. Not as suppliers of stuff, but as makers of meaning. Not as production capacity, but as judgement capacity. Not as people who can simply create outputs, but as people who know what should exist and why.

How Creative Professionals Can Stay Valuable in an AI-Driven Market

Radley Yeldar

The worst response to this moment would be denial. The second worst would be surrender.

Creatives do not stay valuable by pretending AI is useless. They stay valuable by understanding exactly where it is useful, where it is dangerous and where human intelligence still matters more.

Stylianou frames AI as a multiplier rather than a replacement:

“For me, AI is not a replacement for strong creative thinking. It is a multiplier. It can help with research, exploration, drafting, variations, and testing directions. But it should not be making the final creative decisions. That still requires a human brain with taste, context, instinct, and strategy.”

That feels like the healthiest practical position. AI can be useful in the messy front end of exploration. It can help generate routes, organise thinking, test tonal directions, accelerate references, produce rough drafts and reduce the pain of blank-page paralysis. But if the final decision is outsourced to the tool, the creative has abdicated the very thing they are paid for.

The new creative skillset is therefore not simply “learn prompting.” Prompting is useful, but it is not enough. The more valuable skill is knowing what to ask, why to ask it, what to ignore, what to refine, what to reject and when the output is merely plausible rather than good.

Stylianou lays out where the value is moving:

“The creative industry must move up the value chain. Output alone is becoming less defensible. The real value now sits in taste, strategy, positioning, storytelling, originality, brand memory, cultural understanding, knowing what not to make.”

That list could easily become the job description for the next generation of creative professionals.

Taste matters because AI tends to pull towards the middle. Strategy matters because AI cannot truly understand business consequence unless guided by people who do. Positioning matters because distinctiveness is not produced by accident. Storytelling matters because humans still make meaning through conflict, tension, memory and emotion. Brand memory matters because brands are cumulative; every execution either strengthens or weakens the pattern. Cultural understanding matters because timing, tone and subtext are often invisible to the machine. Knowing what not to make matters because abundance without restraint is just noise.

For designers, Keough’s warning is especially important:

“If you want to move the lexicon forward, don't outsource the thinking.”

That should be tattooed somewhere visible. The temptation now is to let the tool do more and more of the early thinking because it feels efficient. But the early thinking is often where the originality lives. The wrong research, the odd reference, the unexpected question, the intuitive leap, the half-formed sketch, the bad idea that mutates into the good one — these are not inefficiencies. They are the process.

For illustrators and visual artists, Howell’s stance points to another path: defend the value of human authorship not with nostalgia, but with commercial logic.

“Ownership matters: You want a visual language that belongs solely to you. A bespoke illustrative style is a brand asset that grows in value. AI output is, by its nature, derivative. As an agency, our job is to protect your brand’s future.”

That is the argument clients need to hear. Not “please support artists because it’s nice,” although that’s not a bad idea either. The stronger commercial argument is that distinctive human-made work can become an ownable asset. Generic AI output cannot protect a brand from sameness because sameness is baked into the model.

McOwen-Banks offers perhaps the most direct warning to clients:

“If you're reading this and you're currently running AI tools in-house without creative expertise in the room — I'd ask you one question. Do you know what good looks like? Not good enough. Actually good.”

That is the question at the centre of the whole AI creative debate. Because AI will produce good enough all day. Cheaply. Quickly. Tirelessly. It will never complain about another round of amends. It will never push back on a weak brief. It will never tell a client their idea is strategically empty. It will never protect a brand from its own worst instincts unless a human has trained the process to do so.

McOwen-Banks continues:

“Because AI will give you good enough all day long. Competently. Cheaply. Tirelessly. And over time, if that's what you're optimising for, your brand will start to look and sound like everyone else's brand — because everyone else is using the same tools, trained on the same data, reaching for the same average.”

That, ultimately, is the cost of average. Not that one piece of work is terrible. But that everything starts to blur. The brand loses edge. The voice loses specificity. The visuals lose ownership. The work becomes more efficient and less memorable. The short-term saving becomes a long-term erosion of distinctiveness.

The End of Good Enough

The Mission Control Communications

AI and creativity are now permanently entangled. There is no going back to a pre-AI creative industry, just as there was no going back after desktop publishing, digital cameras, social media, smartphones or programmatic advertising. The tools are here, they are improving quickly, and the commercial incentives behind them are enormous.

But the lesson of this moment is not that human creativity is doomed. It is that average creative output is losing its protection.

Manchipp’s line still echoes because it is so uncomfortably precise:

“Average used to be a business model. Now it’s a button.”

That does not mean every creative must become a genius. It does mean the middle has become a more dangerous place to hide. The work that survives and thrives will need more edge, more judgement, more authorship, more cultural intelligence and more strategic clarity.

Heryani is right: if average is a commodity, original thinking becomes both harder and more valuable. 

Stylianou is right: the future value of creatives lies not only in making things, but in judging what is worth making. 

McOwen-Banks is right: AI has eaten the middle, but commoditisation is the end of undifferentiated creative value, not creative value itself. 

Howell is right: when brands buy human illustration, they are buying a point of view, not merely an image. 

Keough is right: the tool can show you where the industry already is, but it cannot move the lexicon forward for you.

So perhaps the real question is no longer “Will AI replace creatives?”

It is sharper, stranger and more useful than that.

Can creatives become valuable enough that replacing them with average no longer feels like a saving?

Because AI can generate. It can polish. It can mimic. It can iterate. It can make something that looks like the thing. But it still takes human taste, context, nerve and judgement to decide what matters.

The age of “good enough” is being automated.

The age of genuinely distinctive creative work may only just be beginning.

Header image by Ashley Gray