What AI Is Actually Doing to Human Creativity in 2026

AI & Creativity • Creators • Human + AI

AI and creativity have a complicated relationship. Generative AI can help people produce stronger ideas in some creative tasks, but research also points to a real tradeoff: individual output can improve while the overall pool of ideas becomes less diverse.

That makes the usual question—”Is AI killing creativity?”—a little too simple.

The more useful question is which parts of creativity AI helps, which parts it can weaken, and how creators can use it without letting the model define every idea.

Research doesn’t support either extreme. AI isn’t automatically destroying human creativity, but giving a model control over ideation, execution, and refinement can change where the creative decisions are actually being made.

What Does Research Say About AI and Creativity?

The evidence is more interesting than either side of the debate suggests.

In a series of experiments published in Nature Human Behaviour, participants using ChatGPT generated ideas that researchers rated as more creative than ideas produced without technology or with conventional web search.

Another experiment involving creative writing found a similar benefit. Access to generative AI ideas improved evaluations of stories, with some of the largest gains appearing among writers who initially scored lower on creativity.

But that study also identified a downside: stories produced with AI assistance became more similar to one another.

That creates an important distinction:

A large-scale study comparing 9,198 humans with LLM-generated responses adds even more nuance. Researchers found that human creativity was slightly higher on average in the divergent-thinking task they studied, with humans showing greater variation and stronger performance at the most creative end of the distribution.

So “AI is creative” and “humans are creative” aren’t mutually exclusive statements. The harder question is what happens when the two work together.

How AI Can Enhance Human Creativity

AI Can Expand the Number of Ideas You Explore

One obvious advantage is speed.

A creator can use AI to generate alternative headlines, visual concepts, story directions, product ideas, or approaches to a problem without spending an afternoon developing every possibility manually.

The real value isn’t necessarily using one of those outputs. Seeing alternatives can help you discover a direction you hadn’t considered.

That approach uses AI to expand the idea space rather than letting it establish the entire space for you.

AI Can Make Prototyping Faster

Creators don’t always need AI to invent the idea. Sometimes the bigger advantage is seeing an existing idea sooner.

A rough concept can become a mockup, draft, storyboard, image, code prototype, or alternative layout quickly enough to evaluate whether the idea deserves more work.

That changes the economics of experimentation. More ideas can reach the “let me see if this works” stage before you invest heavily in them.

AI Can Challenge an Idea You Already Have

Generative AI doesn’t have to be used as an idea generator at all.

It can act as a critic:

• What’s predictable about this concept?
• Which part resembles common approaches?
• What assumption am I making?
• What would the opposite version look like?
• How might a different audience interpret it?

Those questions keep the human idea at the center while using AI to push against it.

How AI Can Hurt Creativity

The strongest concern isn’t that using an AI tool instantly makes someone less creative.

It’s that repeated dependence on generated starting points can influence which ideas get considered in the first place.

AI Can Anchor You to Its First Ideas

Suppose you ask an AI assistant for ten video ideas before you’ve thought of any yourself.

Even if you reject most of them, you’ve already seen one version of the problem space. Those suggestions can influence what you think of next.

This is one reason I wouldn’t use AI first for every creative task.

When originality matters, try producing an initial batch of ideas yourself. Bring AI in afterward to critique, combine, challenge, or extend them.

More Creative Output Can Still Mean Less Diversity

This is one of the most important findings in current AI-creativity research.

Generative AI can improve the quality of individual output while making different people’s work more similar.

Research into AI-generated product ideas has found a related pattern: LLMs can produce strong ideas, yet groups of AI-generated ideas may show less diversity than ideas generated by groups of people.

For creators, that’s a practical concern.

If thousands of people ask similar models for blog topics, thumbnails, hooks, story structures, brand names, or visual concepts, the average output can be perfectly competent while still pushing everyone toward familiar territory.

AI Can Turn the Creator Into an Approver

There is a meaningful difference between creating with AI and repeatedly choosing between things AI created.

A workflow can quietly become:

Generate → Pick → Regenerate → Pick → Polish

The human is still making decisions, but much of the raw ideation and execution now begins with the model.

That isn’t automatically bad. Curation itself can involve skill, taste, and judgment.

But creators should recognize the tradeoff. If you want to strengthen your ability to generate original ideas independently, you still need opportunities to practice generating them.

AI Doesn’t Make Creative Skills Obsolete

The original draft argued that creative skills are shifting from execution toward prompting and curation. That’s partly useful, but it goes too far if it implies traditional creative ability matters less now.

Knowing how to write helps you recognize weak writing. Design knowledge helps you judge composition, hierarchy, typography, and visual mistakes. Musical experience gives you more ways to hear what’s interesting or ordinary in generated music.

AI can lower the execution barrier, but lower barriers don’t eliminate the value of expertise.

In many workflows, expertise becomes the thing that helps you decide:

• Which generated option is actually good
• What needs to change
• What feels derivative
• When the AI misunderstood the intent
• What should be discarded entirely
• When the work is finished

The Difference Between AI-Assisted and AI-Led Creativity

This isn’t a scientific classification. It’s a practical way to inspect your own workflow.

Sometimes an AI-led process is exactly what you want. If you’re creating a disposable placeholder graphic, spending an hour protecting your creative identity would be ridiculous.

For work where originality is the value, though, keeping more of the direction on the human side makes sense.

A Better AI Creativity Workflow

For important creative work, I’d structure the process like this:

1. Start Without AI

Write down a few ideas, sketch the concept, or define what you’re trying to communicate before opening the model.

2. Ask AI to Expand the Space

Don’t just ask for more versions of the same idea. Request opposites, unusual constraints, different audiences, or approaches that violate your assumptions.

3. Choose the Direction Yourself

Don’t let the model’s ranking decide which idea matters. Select based on the goal, audience, experience, and taste you bring to the project.

4. Use AI Where It Removes Friction

Let it help with rough drafts, variations, prototypes, repetitive edits, or other mechanical work where speed lets you experiment more.

5. Put Something Back In

Add the part another person couldn’t get from the same prompt: your experience, observation, humor, visual taste, test result, story, opinion, or unusual connection.

A Simple Creativity Test for Your Own AI Workflow

If you’re unsure whether AI is expanding your creativity or narrowing it, try a small experiment.

Choose a real creative task and make two passes.

Look for more than the number of ideas.

Which set surprised you? Which ideas felt predictable? Did the AI-first version pull you toward directions you wouldn’t otherwise have considered—or did it anchor you to obvious ones?

That’s a much more useful personal test than arguing abstractly about whether AI is “creative.”

What AI Still Can’t Supply for a Creator

Generative AI can produce increasingly sophisticated creative output. That doesn’t mean it automatically supplies the context that makes your work worth paying attention to.

Your side of the equation can include:

• First-hand experience
• Personal taste
• Knowledge of your audience
• Real-world observations
• Emotional context
• Original testing
• Intent
• The willingness to reject a technically good idea

Those ingredients become especially important as competent generated content becomes easier to produce.

The objective isn’t to prove a human can manually outperform AI at every individual task. It’s to create something that reflects a reason for existing beyond the model’s ability to generate it.

AI and Creativity FAQ

Is AI killing creativity?

Current research doesn’t support such a broad conclusion. Generative AI has improved creative performance in some experimental tasks, while other research has identified concerns around idea diversity, similarity, and overreliance on generated starting points.

Can AI make people more creative?

It can help in some situations. Studies have found improvements in idea generation and creative writing when participants had access to generative AI. The effect varies by task and user, and stronger individual output doesn’t necessarily mean greater diversity across a group.

Does AI-generated content count as creative?

That depends partly on what you mean by creativity. Researchers measure creativity in different ways, including novelty, usefulness, divergent thinking, and evaluations of completed work. AI systems can perform strongly on some of these measures, but those tests don’t settle philosophical questions about intention, authorship, or human expression.

Will AI make creative work more generic?

That’s a legitimate risk rather than an inevitable outcome. Research has found lower diversity or greater similarity in some collections of AI-assisted or AI-generated ideas. Creators can counter this by bringing independent ideas, varied source material, unusual constraints, personal experience, and stronger human selection into the process.

Should creators use AI?

The useful question isn’t simply whether to use it. Decide what role AI should have. It can be valuable for exploration, critique, prototyping, and execution without requiring you to hand over every creative decision.

Final Thoughts: Is AI Enhancing or Diminishing Creativity?

AI can do both, depending on what we’re measuring and how it’s used.

It can help an individual generate ideas, overcome execution barriers, explore variations, and prototype faster. At the same time, research gives us a reason to watch for convergence: better individual output doesn’t automatically produce a more diverse creative world.

For creators, that tradeoff suggests a fairly simple approach.

Bring something to the machine before asking the machine to bring everything to you.

Start with your idea. Use AI to stretch it, question it, visualize it, or build it faster. Then make the decisions that turn an output into something you actually want to put your name on.