Creativity Community

Pay people for ideas and you will get more ideas, right? A study in this month's Harvard Business Review says the answer is stranger than that.

Researchers tracked 2,376 inventor-years at publicly listed firms in China after those companies introduced cash rewards for inventions. The reward did not simply raise output. It changed the kind of output, depending on who received it.

Specialists, people whose careers run deep in one technical field, responded by producing fewer inventions, but ones that were more novel and technically significant. Generalists, whose work spans several domains, responded by producing more inventions, faster.

Same incentive. Two completely different creative responses.

This fits what motivation research has argued for decades: rewards do not add creativity, they steer it. A cash bonus tells a specialist "make this one count" and tells a generalist "keep them coming." Neither is wrong. The mistake is assuming one dial moves everybody the same way.

For an entrepreneur or a team lead, the practical question is not "should I reward ideas" but "what mix of depth and breadth do I have, and what am I asking each of them to optimize for?"

If you want breakthroughs, protect your specialists' time. If you want a steady stream, unleash the generalists.

Which one are you, and does your reward system know it?

The largest creativity test ever run put more than 100,000 people up against ChatGPT, Claude, Gemini and friends. The AI beat the average human.

That's the headline. The finding underneath it is more useful.

The study (Bellemare-Pépin, Lespinasse, Olson, Bengio, Jerbi and colleagues, Scientific Reports, January 2026) used the Divergent Association Task: name ten words as unrelated as possible. Simple, fast, and strongly correlated with real creative work. On that task, GPT-4 and others outscored the typical person. Then the researchers looked at the most creative half of the humans. Their average beat every model tested. Among the top ten percent, the gap widened further, and it held up on haiku, plot summaries and short stories.

Two details matter for anyone leading a team. First, AI creativity wasn't fixed. Raise the "temperature" or prompt it to play with word origins and the scores climbed. The machine's originality was a function of human direction. Second, "the most creative half" isn't a genetic elite. Divergent thinking is trainable, and most of the people in your building have never been trained.

So the competitive question isn't human versus machine. It's whether your organization is building people who sit above the line the machine just crossed.

What are you doing, on purpose, to move your team into the top half?

The most rigorously tested creativity tool of 2026 has no login, no subscription and no prompt box. It's a walk.

A meta-analysis published in PLoS One this spring pooled 23 studies and 1,036 participants to ask a simple question: does walking change how we think creatively?

The answer was large and specific. Walking produced a big jump in divergent thinking, the ability to generate many novel, useful ideas (effect size d = 0.93, which in plain terms means walkers routinely out-idea sitters). It did nothing measurable for convergent thinking, the ability to narrow down and pick the best answer.

That split is the useful part. Walking isn't a general brain upgrade. It's an idea-generation upgrade.

Design thinking has always asked us to diverge before we converge. This research tells us where to do each. Widen the field on foot. Narrow it at the table.

So the next time your team books a "brainstorm" in a conference room, consider taking the first twenty minutes outside, then coming back in to decide. Same people, same problem, different sequence.

Where did your last good idea actually show up: at a desk, or on the move?

Your first great idea probably wasn't finished when you had it. Someone at your kitchen table finished it with you.

A 2026 study in the Canadian Journal of Administrative Sciences followed 30 technology entrepreneurs and, unusually, 11 people from their inner circles through the earliest stage of idea development. The finding: the creative work of a new venture doesn't happen inside one head. Founders leaned on a small set of trusted people who did two things at once. They supported the idea, and they challenged it. Those conversations modified, clarified and sharpened the ideas before any customer, investor or co-founder ever heard them.

That's a twist on classic network research. We've long credited weak ties with bringing novelty. Here it was strong ties, a spouse, a college friend, a parent, providing the safety to say a half-formed thing out loud and the honesty to push back on it.

Most ideation frameworks skip this stage. Design thinking starts at empathy for the user. But before that there's a quieter loop: a founder testing a rough idea on people who care enough to be honest.

I wrote a chapter on creativity and innovation in Deemed Essential (deemedessential.org), and this is the part I'd add today: protect your circle. Then ask who gets to be that circle for the people you lead.

Who is the first person you run your half-baked ideas past?

We teach creativity like it's a thinking skill. New evidence says it's also a feeling skill.

A 2026 study in Frontiers in Psychology examined design-thinking-based entrepreneurship education through control-value theory — the framework psychologists use to explain how emotions shape learning. The finding: emotional engagement mediates the link between design thinking and the competencies we actually care about — innovation and interpersonal skill. Students didn't grow creative competence just by learning empathize-define-ideate-prototype-test. They grew it when the process made them feel something: curiosity, pride, even productive frustration.

That matches what practitioners see every day. Hand a team a whiteboard and a framework, and you get compliance. Put them face-to-face with a real person struggling with a real problem, and the ideas change texture entirely.

A practical takeaway for anyone leading ideation: before asking "is our process right?", ask "does anyone in this room feel anything?" Real users, real stakes, visible progress. Emotion is the engine; method is just the steering.

When was the last time work made you feel something strong enough to create from?

Your best ideas don't show up while you're working on the problem. Science has a good explanation for why.

Researchers call it incubation. In a recent Scientific Reports study, people who stepped away from a creative writing task and let their minds wander came back measurably more creative — and the benefit was specific to mind wandering, not just any break. Earlier work dubbed it the "shower effect": moderately engaging activities — walking, showering, tending something with your hands — beat both hard focus and total idleness, because they occupy just enough attention to let ideas recombine underneath.

Here's the uncomfortable 2026 twist: researchers now warn that constant stimulation is crowding out exactly this state. Every gap in the day gets filled with a screen, and the brain's default mode network — where incubation lives — never gets the room.

So the practical discipline isn't another brainstorm on the calendar. It's protected wandering: define the problem sharply, work it hard, then deliberately leave it and move — no podcast, no feed. Entrepreneurs I teach are often shocked that the "unproductive" walk was the most productive hour of the week.

When was the last time a real breakthrough hit you mid-task — versus in the shower, the car, or the walk afterward?

Serendipity has a mechanism — and we can finally see it.

A new study in Creativity and Innovation Management followed 159 entrepreneurs and found that design thinking predicts digital innovation, but not directly. The path runs through serendipity. Founders trained to empathize, reframe, and prototype simply notice more lucky accidents — and convert more of them into real innovations. The effect grew stronger in turbulent, fast-changing markets.

Pasteur said chance favors the prepared mind. This is evidence that the preparation can be taught.

Here's the practical riff: ideation isn't a brainstorm meeting on the calendar. It's a standing discipline of exposure and noticing. You prototype early not because the prototype is right, but because it puts you in the path of accidents worth having. Organizations that formalize this — show the rough thing, talk to one more user, follow the odd data point — manufacture their own luck.

The next time something "lucky" happens in your work, look backward. You'll usually find a design choice that made it possible.

What's the best accident your work ever handed you — and what had you done that let you catch it?

A new study just complicated everyone's favorite brainstorming partner.

Researchers publishing in Frontiers in Psychology (2026) compared design students ideating with generative AI versus traditional methods. The surprise: traditional methods produced stronger divergent thinking — more original, wider-ranging ideas — while AI made the process feel more supported, more immersive, more explorable.

That echoes an earlier finding from Doshi and Hauser: AI lifts individual creativity while quietly shrinking the diversity of ideas across a group. Everyone gets a little better — and a little more alike.

For entrepreneurs and educators, I read this as a sequencing problem, not a verdict. Diverge first: sticky notes, bad sketches, wild analogies — the messy human stuff design thinking was built for. Then bring in AI to stretch, combine, and pressure-test what you found. The tool amplifies whatever you hand it. Hand it sameness, it returns polished sameness.

The creative act — noticing what nobody else notices — is still the entrepreneur's edge.

How do you sequence AI in your creative process: before the ideas, after, or not at all?

Nine people, working separately, invented the same toy. All nine named it the Build-a-Breeze Castle.

That's from a study published in Nature. Researchers asked people to invent a toy using a brick and a fan. Those working alone produced ideas that were almost entirely unique. Those brainstorming with ChatGPT produced better individual ideas — sharper, more polished — but 94% of them overlapped conceptually. Nine participants landed on the same castle.

This is the quiet tradeoff in AI-assisted ideation: quality up, variance down. And variance is the raw material of innovation. A portfolio of nine near-identical good ideas is worth less than nine strange ones, because you only need one to be right in a way nobody predicted.

The fix isn't to put the tools down. Follow-up work found homogenization reverses when you change how you prompt — more specific framings, deliberately different personas, people generating alone before they converge. Diverge first, then let the machine sharpen. Not the other way around.

Which makes the old design-thinking discipline more valuable, not less: protect the divergent phase like it's sacred, because that's the part the tool quietly compresses.

Where in your process do people actually think alone before they think together?

Ever wonder what actually happens between "I have an idea" and "we made it real"?

I think every idea travels through four eras — and so does every person and team who works with ideas.

Idea 1.0 is the spark: learning to notice. Most ideas die here, unwritten, because nobody caught them. 2.0 is craft: shaping the raw spark — sketching, drafting, prototyping, making it sturdy enough to show someone. 3.0 is collaboration: the moment an idea stops being "mine" and becomes "ours," and gets better because of it.

And 4.0 is agency: bringing ideas to life at a pace we've never had before — with AI as a creative partner that can draft, render, and test alongside us. Here's the part I hold onto: the imagination stays human. Tools amplify the spark; they don't supply it.

Most of us have a favorite stage — and a stage where our ideas quietly die. Knowing which is which might be the most practical creative skill there is.

So tell me: where do YOUR ideas thrive — the spark, the craft, the collaboration, or the launch? And where do they get stuck?