The Ground Shifted
On March 22, 2023, I got an email from a colleague, world-renowned portrait photographer Jeremy Cowart. It was titled, “Houston, We Have a Problem.” Cowart has photographed the Obamas, the Kardashians, Dolly Parton, Mark Cuban, and Taylor Swift. He’s done passion projects for the United Nations. In the email, he described his blood running cold when he realized he could no longer reliably distinguish between real photographs of real humans and AI-generated portraits. This wasn’t a casual hobbyist having a weird day. This was a world-class professional saying the ground floor of his industry had suddenly and freakishly shifted.
A month later, another colleague pulled a global fire alarm (on purpose). Boris Eldagsen, an edgy German visual artist, won the Sony World Photography Awards with an image called The Electrician. The only problem was that he hadn’t taken it. He had generated it with DALL·E. He made international headlines by standing on stage, tuxedo and all, and rejecting the award—self-disqualifying on purpose. “AI is not a photographer.” He had planned it all along to force the world to pay attention to a problem it didn’t even know it had (yet).
Cowart couldn’t tell. The panels of expert judges who called The Electrician the best photograph on the planet couldn’t tell. These are trained eyes. These are industry authorities. These are the people whose literal job is to tell the difference. And they couldn’t. This inability to tell what’s human from what’s AI is what I have called the differentiation problem—and photography was just the first field to feel it.
Three years later, the differentiation problem is everywhere. Visual design, illustration, film, architecture. Writing of every kind—reporting, journalism, education, technical, legal. Music. Acting. Voice work. Content creation. In my work with AI, I have been particularly worried about these fields because most of my friends are creatives. But looking beyond individual disciplines, “creative fields” have become the canaries in the coal mine for the broader world of human work in the AI age. This matters because if those whose job it is to make things uniquely theirs can’t prove they did, the problem will eventually reach every field dependent on a human being standing behind any work.
Naming
I have often told my friends that creative work is more than a mere sector. It is a species-level activity. When a writer writes, a photographer sees, a musician composes, a teacher teaches—they are doing something distinctly human. At the heart of this is what I call ‘naming’: recognizing, identifying, and ascribing meaning or purpose to something based on individual experience and intent.
In Genesis, creation doesn’t begin with action. God doesn’t wave a wand or clap. God speaks. Creation is a naming event. “God said…”—and reality forms. Creation, at its root, is the result of language. In the Genesis account, the very first job given to human beings was to name all living things. Call it something. Describe it. Distinguish from the others. Show where it fits in the larger order of things. Whoever those first humans were, they must have had incredible intellect to do that work.
“If language creates, and we’ve placed that power in algorithms, we’re doing much more than automating communication. We are automating creation.” —from Human Is the New Vinyl.
Across human experience, what things are called carries enormous significance. Words always matter. In counseling and therapy, words are as important as any clinical tool, any surgical tool. We use words to describe fear. To put handles on trauma and we can see it for what it is and what it’s not. Mental health professionals know how necessary it is to name the things in us that are broken. Because when we name them, they don’t have to live in some dark, amorphous limbo. They become specific. They get a shape. We can come to understand them.
Naming visions of our preferred future is foundational in the framing and operation of human civilization. Laws are ideas put down in words. Treaties are agreements in words about what comes next. Nations are carved out with words and the signatures of women and men who name themselves as framers and backers. All of human life runs on language.
So here is what I mean by naming. To name something is to declare this is what this is — and to be the one who said so. It means intentionally attaching meaning or identity to a piece of work, shaped by the creator’s own insight, experience, and responsibility. For example, a writer names her sentences by imbuing them with her personal style and choices. A photographer names his frame by choosing the moment and the angle. A therapist names pain so it can be addressed. In every case, ‘naming’ means a human asserts: This is what this is, and I am the one saying so.
That is the act that AI cannot perform—and not because AI is weak.
No, AI is incredibly strong. It’s fluent in every human language. It’s read more text than any human ever could. It writes faster, cleaner, and often more clearly than most of us.
But fluency and formation are not the same thing.
To be a namer is not to have an abstract function. A namer is a person, shaped by lived experience, and willing and able to stand behind a specific set of work. A photographer who has lost someone she loved—and then photographs another grieving family, striving to get the light just right. A writer who has lived the disappointment they’re trying to convey. A composer who fell in love and then lost it. A teacher who has watched a thousand students struggle with the same crucial concept. Naming comes from that kind of human experience. And the resulting work carries it forward.
“AI can mimic tone and frame a decent sonnet… but it can’t grieve the loss of a child. It can’t fall in love, feel the knife of betrayal, and translate them into something honest.” —from Human Is the New Vinyl.
Formation can’t be generated. It can only be gained through the process of living: trying, failing, learning. Struggling, getting close, acting, assessing action, and revising.
With AI, we can now produce output at scale without really standing behind it. We can generate a letter in a snap without really meaning it. We can create an image ex nihilo without ever having seen anything. We can slop-cannon all channels of human communication with fluent material that no human is willing—or able—to claim.
That is exactly what is happening. Everywhere. All at once.
What AI produces is fluency without formation. Output without a namer. And seas of empty artifacts that nobody is standing behind.
This is the true crisis: not just technological disruption, but the erosion of human accountability in our creative and productive processes.
The Question
In searching for solutions, I am in legislative think tanks. I am in C2PA rooms with Microsoft, Adobe, Sony, and the BBC three or four times a week. What I have noticed is that all the responses we have mounted so far try to authenticate fluency. Copyright law tries to claim the namer after the fact. Detection tools try to reveal the namer from the outside. Watermarks and digital signatures try to embed a namer into the file. Yet all of them fall short.
Because naming doesn’t come from a statistical point in a dataset. It is a uniquely human behavior.
And you can’t build a system that performs a human act on behalf of a human who has stopped performing it.
Which leaves us with a question our civilization hasn’t had to answer before.
What does it mean to be a productive human when output no longer requires a human?
What does it mean to create when creation no longer requires a creator?
What does it mean to name when the thing being named was produced by something that cannot itself be named?
These are not rhetorical questions. They are live ones. They are being answered right now, in real time, by everyone who posts, publishes, signs, or claims anything. And most of us are answering them without realizing it.
And being unaware of danger is always the most dangerous place to be.
Faith communities, more than any others, should recognize this terrain. The question of what makes a human word trustworthy—what makes a promise binding, what makes a witness real—is not new. The vocabulary already exists. Covenant. Oath. Testimony. Witness.
Consider the courtroom oath. “I swear to tell the truth, the whole truth, and nothing but the truth.” The words don’t make a person honest—a liar can say these words and still lie. But the oath does something important. It marks the moment. It draws a line. And when trust breaks down later, we can point back to exactly where it was promised and say, “Hey, remember when you said this? What gives?”
Whether we use that vocabulary, or let it fade (or be drowned out by the overwhelming barrage of digital noise) while we argue about technical fixes, is a choice we are making—whether we notice or not.
We are still seeking clear answers.
What is clear, however, is that the question is real—and we are already late in asking it.
Remember, the ground shifted in 2023.
Yet three years in, most of the conversation is still about jobs.
Addendum: A Note on What I’m Building
In April 2023, a month after that email, I founded VerifiedHuman — a global certification platform for creators whose work is substantially their own.
Our mark signals only this: I made this work, and I stand behind it. It is not a declaration of superiority, but a transparent commitment. Certification is based on trust, not detection—a clear answer to the confusion about agency and authenticity in the age of AI.
We have members in 25+ countries on six continents: writers, visual artists, musicians, voice actors, educators, organizations, advocates, and content creators.
I recognize this isn’t the only approach. And it may not ultimately be the right one.
Still, we couldn’t do nothing — because these questions are too important to ignore.
This essay is a short version of the arguments I explore in longer form in my book, Human Is the New Vinyl.
Views and opinions expressed by authors and editors are their own and do not necessarily reflect the view of AI and Faith or any of its leadership.
