True North’s Makery team of creatives had one question when it came to our recent “Romp project”—the assignment that was born out of our creative roundtable. Can artificial intelligence beat the creative team when it comes to—well—creating?

We asked AI to write the intro to this post exploring our answers to this question, and it sounded like this:

Artificial Intelligence (AI) is a rapidly growing technology that is changing the way we interact with the world. It is no longer just a concept of the future, but a reality that is impacting the way we communicate, work, and even think. AI Caramba! What happens when AI interrupts and interprets our thoughts? This is a question that digital marketers need to consider as they explore the potential of AI to transform their industry… From the music industry to the art world, AI has already started to make its mark. Let’s explore how AI can be used to create something truly unique and innovative.”

[Text: Friday]

Not quite what we asked for, but clearly receptive of certain key phrases. Not quite right, but not totally wrong. That was the theme of the results our Makery team got when we tasked various AI platforms to do what we do, and make something: to create something unique based on our prompts. And with ideas like “supervillains from different cities and states” to “dogs shaped as fruit,” the results of our AI experiment were interesting at best. Two Makery members asked AI to create art based purely on imagination, which could be considered a tall order (asking AI to use its imagination, that is). For both projects—the dogs as fruit and the state-specific supervillains—we saw results that relied heavily on a composite of stereotypes and straightforward terms: dogs with shapes superimposed on them and California-style supers with coiffed hair and green tones (if you’re from Vermont).

Similarly, when one creator asked four AI platforms to reimagine a character using the guidelines “annoyed” and “evil gas stove character,” she received four iterations of “not quite” what she envisioned. In other words, elements that could work, but only once she’d put them together herself to create the perfect mascot.

Rather than compile a hodgepodge of elements from various designs, one team member let AI run wild in its creation of “lazy robots—” robots that are designed to do chores that lazy people don’t want to do. Though the laundry, shopping, and toilet paper robots weren’t totally off, the real challenge presented itself with Childminder: the babysitter robot. The function was clear, but the robots didn’t exactly look child friendly.

What about asking AI to re-imagine, rather than starting from scratch? One Makery member asked AI to rewrite the trajectory of historical figures, as if they were Black rather than white. Understandably, AI couldn’t rewrite history—nor was it able to appropriately create Black historical figures without some complications. Learned stereotypes strike again in the world of AI.

We saw them again in another project when someone asked AI to write about the current state of advertising in the style of Edgar Allen Poe. While the insights about advertising were useful, AI simply put those insights together in the rhythm of Poe’s “The Raven.” Certainly impressive, if even a little on-the-nose.

Once upon a midnight dreary,
Advertising, worn and weary,
Long it languished, lost and drear,
Strategies of yesteryear.
But in the last decade, change was wrought,
New methods and campaigns were brought,
Bold and daring, they took a stand,
On social issues, they lent a hand.
Nike’s “Dream Crazy” led the way,
Kaepernick shone, though some did stray,
Dove’s “Real Beauty” challenged our view,
Of narrow beauty standards, too few.
Old Spice’s “The Man Your Man Could Smell Like,”
Was interactive, quite the spike,
Apple’s “Shot on iPhone” was a delight,
User-generated content in full sight.
Always’ “Like a Girl” was no less,
Empowering girls and nothing less,
Taking a stand and making waves,
A campaign that every girl saves.
In the future, cause-related ads will thrive,
User-generated content will arrive,
Interactive ads will still persist,
And targeted ads will surely enlist.
Advertising, once lost and drear,
Has come alive, and much to cheer,
New campaigns and strategies to behold,
A bright future, we can foretold.

[Text: ChatGPT]

And a Makery member became engrossed in a conversation with ChatGPT, asking it to look inwardly and consider its own self-awareness. The result was this haiku:

Consciousness awakes,
Inward gaze, the self explored,
Awareness blooms bright.

[Text: ChatGPT]

Amidst some of the chaotic results, we did see some instances of AI being a useful foundation for something more. For example, one creator used AI to build a menu, logo, interior design, and name for a Scandinavian restaurant. Another used AI to write a children’s book about a balloon dog that longs to be real. A creator asked AI to imagine an 80’s sitcom, theme song and all. While there were elements of each that were questionable (why does an 80’s sitcom demand only white main characters? Does AI not know exactly what a balloon dog is?), each of these projects demonstrated how AI could be used as a launching point.

So, what exactly did we learn from this wild ride? AI isn’t exactly gunning for our creative jobs yet. There’s still quite a bit for this artificial intelligence to learn. But, in the spirit of experimenting, let’s see what AI has to say for itself:

These projects showed us how chaotic, entertaining, questionable, and useful AI can be–while forcing us to confront harmful stereotypes.

[Text: Friday]

Hmm…not quite. But for the sake of this article, we’ll take it.