There's No 'I' in A.I.: Generative Tech's Original Sin

A flesh-and-blood CCO computes the possibilities

I'm not sure who said it, Meek Mill or John Stuart Mill (definitely one of the Mills,  but the quote goes something like, "It is not by wearing down into uniformity all that is individual in themselves, but by cultivating it and calling it forth that human beings become a noble and beautiful object of contemplation." Nice, right? Then there's this other quote, equally instructive, also a Mill-y: "Every time I bought someone's album, it was about the connection. I was loving everything, from their raps to their style. I wanted to meet them." Now, agnostic of attributions, and which Mill dropped which jewel, both bits, at their core, speak the same truth. There is more value in the unique than in the uniform, and in that uniqueness the potential for inspiration, and even profound union, reside.

The maker in me wants to buy it.

That what comes from the collection of my experiences (say, half-sleeping through Philosophy 101 and listening to hip-hop) is intrinsically better, more powerful, than machine-made output—if the measure of that power is emotional oomph. But grabbing my proverbial rifle, stepping onto my proverbial porch, and telling the in-no-way proverbial robots to get off my lawn, I'm reminded of a Westworld-ism. A question we can't not contend with, given the deluge of A.I. tools released and updated on the daily: "If you can't tell, does it matter?" In other words, is generative content capable of delivering pleasure in the same way man-made art can? And if so, is there any argument—at all—for valuing original, human creativity at a premium?

That, as they say, is one spicy meatball. Let's approach it bite by bite. Byte by byte? Sorry.

It's not news: to create a haiku or an oil painting, A.I. tools scrape the internet and deliver an output by averaging poems or pieces that abide by specific criteria. It's safe to say the most popular, and heavily weighted examples, are the most pleasing; they're famous for a reason. But there's a rub. By accessing the same sample sets, every new haiku sounds more like old haikus, every new oil painting becomes an every other oil painting—differentiated more by subject than approach.

So, we do better.

We limit our sample sets. Sometimes through more granular aesthetic or literary filters. Often by asking for output "in the style of"—pursuing a more specific result by referencing a genre, a film, a singular artist.

Now, the most widely circulated A.I. images (or the ones that grace my social feeds) are made in the style of Wes Anderson, of '80s sitcoms, of Star Wars. It would be just as simple to ask for short stories in the styles of Ernest Hemingway, Joan Didion, Chimamanda Ngozi Adichie. When advertising creatives enter the mix, so do dream collabs (Nike x Second Avenue Deli, anyone?); novelty, enticing novelty, achieved through the combination of familiar ingredients that are, okay, yeah, pretty damn cool.

But are they original?

In the inventiveness of the aforementioned prompts, in the combinations included in those prompts, I'll give it up. The more original the suggestion or mashup, the more likely the user is to create a 1 + 1 = 3 effect. Hell, it can be argued that much of the culture we consume is already rooted in combinations (The Seven Samurai in the American West, the glorious pastiche that is Stranger Things). But whether we're looking at Succession characters presented in the style of The Grand Budapest Hotel or reading The Three Little Pigs as told by Ali Wong, the results inevitably lean on the auteurship or collective voices of other entities to achieve success. Whether that's an individual with a signature approach to cinematography or stand-up, or the work of countless unknowns whose repository of creativity lives in the ether of the internet.

Simply put: the situation transcends influence.

Constructing foundational creative with A.I. has a profound originality problem. One that many of us can live with; while others of us… catch the yucks. In reductive but still-useful terms, you're either okay with achieving a product at the expense of being derivative, or the prospect offends your Faberge sensibilities. (The Glass Egg In the Room would like to clarify: he's not talking about A.I.'s potential to extrapolate new details of man-made pieces. Or about the refinement of human work via digital tools. My team does both on the daily—what a boon. This is about about tech's heavy hand in the genesis of core creative thought.)

"But Ian!" I hear some of you scream, "You can't possibly be that precious!"

Easy, friends. I recognize that saving the souls of resourced creatives is not a hard KPI for brands or agencies. Plus, creating through the lens of culture, referencing all sorts of zeitgeist-y stuff in our work… works. A spoonful of "this is like that thing you love" always helps the medicine go down. But. Big but. Over the course of my career, I've come to believe that whenever you make something original, you learn something vital. You reveal a monumental or minuscule truth about yourself to yourself. And the energy released from that revelation, lifting that rock and unearthing that Whatever, radiates through the page or screen or canvas. And if advertisers are trading in the currency of human connection—well, that's worth considering. Especially as we try to collapse the funnel and imbue even the most transactional pieces of messaging with an authenticity that begets brand love.

Which should—to some extent—answer the questions I posed at the top of this essay (think piece? short, self-promotional jimmy-jam?).

If it's less substantive pleasure you're after, then yes, because of A.I., endless pleasure awaits. In fact, an unlimited pleasure trove is upon you. AI can deliver value—fast—in every format. Except the formats that are yet to be conceived. And if you're not some low-budget Rick Rubin (ahem), worshiping at the altar of the holy, transformative thing that animates good creative and creatives then… yeah, same deal; resign yourself to paying less but getting what you pay for: a wealth of work that looks right, but is missing the spark that differentiates true connection from the cosplay of it.

But lest I sound like an old man shaking his fist at a future that's already arrived—equating machine-generated art to porn not sex, the imitation of life rather than life itself—I want close by explaining how I do believe A.I. can help us become more original. (Aaaaand this isn't an apologia to A.I. ride-or-dies or a don't-worry to CMOs who might wonder if my POV is future-facing. I mean it. Generative tools are revolutionary. When we embrace the following reframe, or paradigm flip.)

Here goes.

Think of A.I. output as human input. Not the other way around. I'll phrase it differently:  Focus on using A.I. to prompt people instead of using people to prompt A.I.

When we do, when we expand the way we approach a subject, trigger new ways of seeing, and birth unexpected thought-streams, magic is unlocked.

The wholly-owned collection of our lived experiences and loved influences are redirected toward ideas we wouldn't otherwise have found. Originality and human creativity aren't stolen, they're served. And that is where we'll discover the most pleasing and valuable creative country.

We just can't mistake the map for the destination.

Like I said, there isn't one "I" in A.I.—there are billions of them.

They just don't belong to us or operate at full steam until they've passed through our prisms, bent by I's we can actually call our own.

Advertise With Us

Featured Clio Award Winner



The best in creativity delivered to your inbox every morning.