The Essence of Creativity and How Generative AI Can Help Unlock It
What research says about creativity and how you can use GenAI to access it.
Ground-breaking research from 1971 suggests that generative AI might be more essential to creativity than you think.
I’ll explain the grapes later.
Research revealed the elements of creativity in 1971
In 1971, a paper entitled Discovery-Oriented Behavior and the Originality of Creative Products by Mihaly Csikszentmihalyi and Jacob Getzels based on research at the University of Chicago described a study of 31 art students.
Each subject was given 27 objects on a table. Among them was an antique book, a camera lens, a floppy velvet hat, a polished gear shaft, and a bunch of grapes.
Students were asked to build their composition on a second table using these objects, then draw it.
No time limit was given.
Observers scored six behaviors, three during problem formulation and three during problem solution.
During problem formulation, they measured:
The number of objects manipulated. How many objects each artist touched before beginning to draw.
Uniqueness of objects chosen. Uniqueness of objects selected. For example, the book was the most commonly selected item and the camera lens was least popular and got the highest score.
Discovery-oriented behavior. A one was given if the artist simply picked an object and used it. A five was given when the they felt its weight and texture and experimented with it — for example, folding the hat into different shapes, looking through the lens, and playing with mechanical parts.
During drawing, scores were given as follows:
Openness of problem structure. The longer the subject sketched and played with the drawing before creating the outline with all its final elements, the higher the score.
Discovery-oriented behavior while drawing. A one was given for drawing without interruption, a two if the artist switched from one medium to another, a three if they rearranged their objects, and a five was given if the artists changed medium and re-arranged objects.
Changes in problem structure and content. A one was given if the subject copied the arrangement, a two, three or four was given for various changes in perspective or position, and a five was received if a nonexistent object was added.
Well-known artists and art critics rated the 31 drawings on three measures:
Craftsmanship, or technical skill,
Originality, or imaginativeness, regardless of craftsmanship, and
Aesthetic value — how would judges rate the work in a competition?1
The Findings: What Creative People Do
The most valued and original work came from artists who:
Considered the most elements before structuring the problem,
Used unusual elements as the focus of the problem or
If a more usual element is selected, they thoroughly explore it through multiple sensory channels before using it.
During drawing, the best results came from artists who:
Didn’t settle on structure too early,
Tried various methods, old and new,
Transformed their objects into a new form.
As Einstein said, the key was problem formulation, not the craft of making the art.
Generative AI Is a creativity coplilot
Generative AI is a reformulator, a brainstormer, a copilot for exploration. It helps you consider more ideas, uncover unusual ideas, explore old ideas from new angles, consider new presentation structures, try various forms of presenting ideas, and recombine interpretations into a new, meaningful form.
To illustrate this point, I redid one of my favorite exercises I described in The Joy of Generative AI Cocreation generative AI. It’s a digital version that uses the findings of Csikszentmihalyi exercise: pick an idea you want to express (in my case, a simple object, grapes), and work with a tool (in my case, the Jasper AI engine) to generate as many ideas you can from different angles, styles, and perspectives.
In an hour, I tried over 200 ways of looking at grapes. My favorites, above, were:
Grapes With a Macro Lens (a new angle — close),
Grapes By Van Gogh (an old medium)
Grapes on Mars (in a new context)
Grapes Held by W.E.B. Du Bois (I liked AI’s version of him holding the grapes)
Grapes as a Hot Air Balloon (recombining grapes and the Pixar movie Up)
Grapes in the Sketching Style of Saul Bass (a new medium)
My favorite, and the one I chose for the cover image of this article, is “Grapes on the Cover of Vogue.” AI’s result shocked me by using grapes to ordain a digital model with eye contact that might make Anna Wintour say yes.
Am I an artist? Certainly not.
Am I creative? If the definition of creativity is how one imagines new elements to use, combines ideas, and reformulates ways to express themselves, I say yes.
Read The Joy of Generative AI Cocreation about how to use AI during the creative process, and It’s Not About the Camera.
Other Substacks I Love
Correlation of .80 on average,
The distinction between generative 'creativity' and art is a subject of tremendous weight at this moment in our evolution.
That's why I wrote this piece, using the intersection between fluffy sheep, satanism and the metaphysical substance of 'sentience.'
Sheep Against Satan!
(Fifth Generation Warfare for the Epistemologically Challenged)
https://walkingwithgoats.substack.com/p/sheep-against-satan
It reminds me of the game where you take an every day object and see how many uses you can come up with.
A lot of people get stuck in the definition of the objects intended use and can't go further.
I always found it was helpful to think of the most absurd uses and then work from there. You'd find some pretty useful ones that way.