AI Can Already Do Part of Your Creative Work. Now What?

You open an AI tool and ask it for something that, a few years ago, would have taken a chunk of your day to make.

An image.

Then another.

Then twelve variations.

Seconds later, you have several visual directions in front of you. Some are shaky. Others are genuinely usable. And with every new generation of models, the difference gets a little harder to ignore.

Your first reaction might be a small rush:

Wow. This is going to save so much time.

Then comes a second thought. Much less comfortable.

If this part of my work becomes so easy to get… what still makes me valuable?

AI is already more than a tool for removing backgrounds or correcting colour. It writes, illustrates, edits, codes, suggests concepts, generates variations and is starting to connect several steps in the same workflow.

That doesn’t mean it does everything. It certainly doesn’t mean it always does it well.

But it’s enough to change the question.

For a long time, part of a creative professional’s value came from being able to do something not everyone could do.

Mastering Photoshop. Building a composition. Editing video. Retouching a photograph. Turning a brief into a visual identity.

What happens when some of those capabilities become available in a few clicks?

The easy answer would be: it’s fine. Humans will always have the ideas, and machines will execute them.

The trouble is, that boundary is already starting to disappear too.

And that’s where the conversation gets interesting.

The machine is already doing more than executing

In April 2026, Adobe surveyed 1,433 professionals and creators in the United States and the United Kingdom about their use of artificial intelligence. Brainstorming and ideation were already among the uses studied.

That doesn’t tell us whether the ideas were brilliant. The study relies on self-reported practices and perceptions. But it does challenge a comforting picture of AI: one where machines handle repetitive tasks while humans keep all the creative thinking to themselves. [IA1]

The tools themselves are moving in that direction.

In June 2026, Adobe announced assistants capable of linking several operations within some of its applications: organizing media, preparing an initial edit, adapting files…

In other words, we’re no longer talking only about a machine that makes things faster.

We’re beginning to talk about a machine that also helps work out what should be made.

If that’s the case, simply saying “creatives will always have the ideas” is no longer enough.

We need to look a little further for our value.

When knowing how isn’t enough

This is probably where things start to feel uncomfortable.

A creative profession isn’t just a brilliant idea arriving at the right moment.

It’s also years spent learning to see.

Understanding why a composition doesn’t work. Why lighting makes an image feel flat. Why an edit lacks rhythm. Why a typeface that looked perfect on screen suddenly becomes illegible on a poster.

Over time, some decisions become almost instinctive.

When a tool starts producing, in seconds, part of what took you years to learn, it’s natural to wonder:

is all that knowledge still worth anything?

The short answer is yes.

Perhaps just not in quite the same way.

A skill can remain useful while becoming too widely available to justify the entire value of a profession on its own.

Knowing how to use Photoshop is still useful.

So is knowing how to edit video.

But as execution becomes more accessible, the question changes:

what were we actually doing beyond execution?

And perhaps we’ve been putting too many different things into one word: creativity.

Take a simple project.

A music school wants to attract more adults who have always wanted to learn but have never dared to start.

Before thinking about a poster, a video or a campaign, we need to understand what’s holding them back.

Is it the price?

A lack of time?

The fear that they’re too old to begin?

Or simply the rather unpleasant prospect of being bad at something in front of other people?

The answer changes what we create.

And that’s when we realize a creative project actually has several layers.

First, there’s strategy: understanding the problem we’re really trying to solve.

Then the idea: finding an interesting way to respond.

There’s taste: the references we choose, what feels right, what we reject because it doesn’t belong to the brand or the world we want to build.

Then comes art direction: making sure the image, lighting, words, typography, rhythm and movement all express the same intention.

Only then comes execution: photographing, drawing, editing, designing, coding, retouching.

Behind that sits production: organizing files, formats, approvals, versions and deliveries.

And above it all, there’s one last thing we tend to forget:

judgment.

Looking at what’s been made and asking:

A piece of creative work can be beautiful and completely useless.

It can be technically flawless and send the wrong message.

It can even be original… for all the wrong reasons.

AI can take part in almost every one of these layers.

It can suggest ideas, propose directions, create variations, help with writing, generate images and organize files.

That’s precisely why drawing a clean line between “human” and “machine” would be a little too convenient.

The reality will probably be much blurrier.

The International Labour Organization and NASK describe AI exposure in terms of tasks, rather than the automatic disappearance of professions. A profession can change profoundly without being eliminated. [IA3]

Which brings us to an important distinction:

A graphic designer doesn’t just sell a file.

A photographer doesn’t just sell an image.

An art director doesn’t just sell a beautiful Pinterest reference.

They also make decisions, interpret context and choose what deserves to stay.

And sometimes, they recognize something much harder to spot:

a good answer to the wrong problem.

The problem with the good idea everyone gets

This is where AI becomes particularly interesting.

It can produce surprisingly good things.

But “good” doesn’t always mean “distinctive”.

A study published in 2024 by Anil Doshi and Oliver Hauser observed a fascinating pattern: in their short-story writing experiment, access to AI-generated ideas improved some individual evaluations… while making the stories more similar to one another. [IA4]

In other words:

each person could receive a better idea while contributing to a less diverse whole.

Of course, one short-story experiment can’t predict what will happen to every visual identity, campaign or illustration.

But it raises a question that already feels familiar in creative work.

Imagine an entire market.

Every vendor visits the same adviser and asks:

The adviser is very good.

They analyse what works.

They suggest a current palette, an interesting typeface and a few irresistible references.

Everyone leaves delighted.

The next morning, the market opens.

Every tablecloth is terracotta.

A caricature, obviously.

But the risk is real.

If we use the same tools, drawing on similar references, to solve problems framed in the same way, we can end up with plenty of “good” proposals that begin to look remarkably alike.

Perhaps this is where another kind of creative value emerges:

knowing what to bring to the machine before asking it for anything.

Originality doesn’t necessarily begin in the prompt.

It can begin much earlier.

In a client conversation. In something observed on the ground. In an expression people actually use, an unusual constraint, an archive, a specific cultural reference, or an experience someone outside the project would never have thought to seek out.

Let’s return to our music school.

If it simply asks:

“Create a warm, inspiring campaign for adult beginners”

it risks getting exactly what we’d expect.

Beautiful instruments, smiling people, soft lighting, lines about following your dreams.

Nothing disastrous.

Now imagine that, after talking to a few hesitant adults, the team discovers something else:

they aren’t afraid of learning. They’re afraid of looking ridiculous while they learn.

Everything changes.

The campaign no longer needs to show mastery.

It needs to show that not having mastered something yet is allowed.

Suddenly, a technically perfect image of an adult playing the piano beautifully can become… the wrong image.

Not because it’s poorly made.

Because it answers the wrong problem.

That’s where judgment really comes in.

Not to decide which image is prettiest.

But to decide which one best serves what we’re trying to do.

So where does our value lie?

If AI can contribute to execution, suggest ideas and generate dozens of variations, it’s tempting to search for the last thing it will never be able to do.

I don’t think that’s the right question.

A creative professional’s value may lie less in some mysterious, exclusively human ability than in how several skills work together: understanding context, building a direction and judging what deserves to be pursued.

Take taste.

Having taste isn’t simply looking at an image and finding it beautiful. It’s understanding why it works here, for this brand, this audience and this particular problem.

A direction can be gorgeous and still be wrong.

Think of our music school again. AI can generate a perfectly warm campaign with beautiful instruments, smiling adults and soft lighting. But if prospective students aren’t afraid of learning and are mostly afraid of looking ridiculous while they learn, the entire campaign needs to change.

That insight doesn’t automatically appear in a prompt. First, someone has to speak to the right people, observe the problem and ask the right questions.

Perhaps that’s where value is starting to move:

before the answer, into the quality of what we’re trying to understand.

When making gets easier, choosing gets harder

Being able to generate fifty visual directions in an hour doesn’t necessarily bring fifty times more clarity.

Mostly, it brings a lot more decisions.

Which one actually addresses the problem? Which one fits the brand? Which one will work beyond a beautiful image? Which one just looks like whatever happens to be trending?

Judgment means building criteria solid enough to answer those questions.

Sometimes it also means saying no: to a beautiful image that sends the wrong message, an original but useless idea, an animation added simply because it’s possible, or a direction that could belong just as easily to five competitors.

As possibilities multiply, knowing what not to keep matters more.

Giving it all a direction

This is also what makes art direction interesting in a world full of tools that can produce.

A campaign isn’t a collection of beautiful things. The image, words, casting, colour, typography, movement and rhythm need to express the same intention.

AI can speed up the making of each element.

But someone still needs to be able to say:

“We want the person who’s hesitant to start to recognize themselves in this scene before they even read the copy.”

From there, the choices become much more precise. We can look at the casting, posture, lighting and tone of the words and ask: is all of this really saying the same thing?

Art direction stops being an aesthetic layer added at the end.

It becomes what gives the intention a coherent form.

Being more valuable isn’t enough. People need to see that value.

Understanding context, building a direction and knowing how to choose are useful. But there’s a very practical problem: much of that work is invisible to the client.

The client sees the logo.

They don’t necessarily see the three directions abandoned because they wouldn’t work on mobile, the conversation that exposed a positioning problem, the references ruled out because they looked too much like the competition, or the thirty minutes spent figuring out why an otherwise lovely idea said exactly the wrong thing.

If that work stays invisible, it becomes easy to compare your service with what a tool produces in seconds.

The challenge for creative professionals isn’t only to develop their value.

It’s to learn how to show it.

1. Start by understanding the problem better, not by producing a solution faster

Before Photoshop, Figma, Midjourney or After Effects, there’s a skill that’s becoming harder to replace: knowing what to look for.

Ask more questions before you begin. What is the client really trying to change? What does their audience understand today? Where is the friction? Which constraints are non-negotiable? What has already been tried?

A brief isn’t just a list of deliverables.

It’s a hypothesis about a problem.

The more precise that hypothesis, the more interesting the material you give your tools — human or artificial.

It’s also something the client can understand.

Instead of opening a presentation with:

“Here are three logo options.”

start with:

“Here’s what we understood about the problem, what we want people to perceive, and the criteria we used to develop these proposals.”

The conversation changes immediately.

You’re no longer just presenting files.

You’re presenting decisions.

2. Train your judgment, not just your tool skills

Learning new tools still matters. Ignoring AI on principle probably won’t make anyone more competitive.

But learning only how to get beautiful results from those tools won’t be enough either.

You need to learn to critique what they produce.

Take a proposal and make yourself explain exactly what works and what doesn’t. Not “I don’t like it”, but: the hierarchy draws the eye to the wrong place; this image changes how the product is perceived; this tone feels more luxurious but less approachable; this composition works on a poster but falls apart on mobile.

Judgment also grows from looking at things outside your own discipline: photography, cinema, architecture, music, publishing, fashion, advertising, popular culture.

AI can expand your library of possibilities. It doesn’t remove the need to learn how to read what it gives you.

3. Don’t just sell the result. Show the reasoning.

This may be one of the most important shifts for creative professionals.

If your portfolio shows only:

logo → mockup → poster → thanks for looking

a client can reasonably conclude that your work mainly consists of making those objects.

Show more of the problem, the constraints, the choices and their consequences.

Why this direction rather than another? Why this typeface? Why did you remove something? How does the system work in different situations? What did you want the audience to feel or understand?

Even a small case study can follow a simple sequence:

problem → observation → decision → execution → result.

That doesn’t mean publishing every draft or turning every project into a thesis.

It means making visible what the finished work conceals.

Because that’s where a client can begin to understand the difference between:

“someone made a beautiful image”

and

“someone understood my problem and built a response that fits”.

4. Use AI to increase your value, not just your volume

The obvious impulse is to use a faster tool to produce more.

Ten visuals instead of three. Twenty concepts instead of five. Four times as much content.

But more isn’t always more useful.

The time saved can go elsewhere: exploring genuinely different directions, preparing a better presentation, checking more formats, doing research, trying out an idea, improving an interaction, documenting decisions.

If AI saves you two hours, you don’t have to fill those two hours with more execution.

You can invest them in what improves the quality of the project.

There’s also a tempting trap worth resisting: making your ability to use AI your entire professional identity.

Tools will change.

Models will change.

The ability to understand a problem, build a direction and evaluate a solution will probably outlast the next update more comfortably.

5. Speak to clients in the language of their problem

A client doesn’t necessarily wake up thinking:

“I need excellent art direction.”

They’re more likely to think:

“People don’t understand what we sell”, “our business looks less credible than it is”, “our launch is going unnoticed”, “our image no longer reflects what we’ve become”.

That’s where your expertise needs to connect.

Instead of selling only:

visual identity, photography, website, campaign

explain what those things help solve.

Without promising impossible results.

By clearly connecting your work to the problem.

An identity can improve a brand’s consistency.

Better website architecture can make an offer easier to understand.

Art direction can stop every piece of communication from telling a different story.

Your client doesn’t need to become an art director to understand your value.

It’s your job to translate your expertise into a language that belongs to them.

6. Accept that speed is no longer proof of value

This may be one of the most disorienting changes.

For a long time, some skills were valuable partly because they took a lot of time.

Now, some steps can take minutes.

That doesn’t automatically mean the whole project is worth less.

If a tool lets you produce an initial exploration in twenty minutes instead of two hours, the client isn’t necessarily paying you for the hundred minutes that disappeared. They’re also paying for your ability to know what to ask, recognize mistakes, take responsibility for the result and deliver something usable.

But you can’t simply declare that value.

It has to be apparent in how you work: your brief, presentations, questions, case studies, explanations and the quality of the finished work.

The faster execution gets, the less convincing “this took me a long time” becomes as a justification.

And the more important “here’s why this is the right solution” becomes.

Now what?

Perhaps the mistake would be to spend the next few years trying to prove that AI will never do what we do.

It will probably do things we thought impossible not long ago.

It will perform some tasks better.

Others less well.

And it will probably do some in ways we haven’t imagined yet.

The useful question isn’t where to draw an eternal boundary between human and machine.

It’s whether we can look clearly enough at our profession to understand where our value moves when the tools change.

For a graphic designer, illustrator, photographer, editor or art director, that may mean becoming better at observing, questioning, connecting, choosing, explaining and directing — while using new tools well enough to avoid wasting time on what they can already speed up.

Perhaps above all, it means learning to make that work visible.

Tomorrow, your client may generate twenty logos, ten images or three campaign proposals before your first meeting.

What they’ll still want from you probably won’t be just a twenty-first option.

They’ll want someone who can look at all those possibilities and say:

this one addresses the problem, here’s why — and here’s how we’ll turn it into something that actually works.

AI can produce more and more.

Our work may increasingly be about getting better at choosing what deserves to exist.