How do we train AI to recognize what is a human? In other words, what is an instance of the human? In my non-technical, lay-person’s understanding, AI makes this distinction by differentiating between semantic segmentation and instance segmentation.[1] In a semantic model, the output would be a silhouette of the whole group against the background. In the instance segmentation model, however, individuals can be recognized. Below, you’ll see an example of Semantic Segmentation (left image) with a group in silhouette, and Instance Segmentation (right image), where individuals are recognizable.[2]
I recently drew upon this differentiation in a poem I wrote entitled, “Instance of Human.”
INSTANCE OF HUMAN
Today, in a universe of artificial intelligence,
I am an instance of human
even if other life forms
exist in the same photograph. A willow is not a human.
Neither a patch of sky nor the full moon.
My right hand, captured in a box,
is parsed from the body yet a part of it,
a flesh-colored blush to the fiddleneck
length of my forearm.
A shock of hair falling
across the eyes might be another instance of human,
or the late afternoon light
falling across my face as it has a thousand times
in half a century. The sunset is a sum
of crepuscular parts, a segmentation
from dusk to dark. To be human is
my ear against a summer oak,
a choir of double-bloom cicadas,
almost mechanical yet very much alive.
Over a lifetime, how many instances of human
have you imparted to this world?
Is that a mourning dove in flight
or your hand sweeping the magnolia buds?
Are you awake, my friend
or is this a weepy fern bed in the woods?
Did you leave this rake
by a heap of speared sorrows?
How are the fugitive, the migrant, and the undocumented
also instances of human?
If we ever had to prove
we are more than a series of keypoints,
than a variation of poses in a span of a lifetime,
than a flickering group of silhouettes—
our instances of the human yield far more than
one person, one soul
as an object in an image.
This poem explores, through sequential imagery and rhetorical questions, our instance of human—we are more than data or a series of keypoints. The poem also considers how a person– an instance of a human—is more than the sum of technical actions, but rather, a holistic synthesis of lived experiences, memories, testimonies, and yes, alchemy of emotions and intelligence that are more than one identifiable person or an object. A biblical takeaway might be that as humans we are created beings, created in God’s own image as imago Dei, and as image-bearers of God, exist as God’s living poems or poiema, handiwork, workmanship, or masterpiece. As the poetJohn Poch has likewise written in his book, God’s Poems, poiema refers to a “a made thing,” in translations of the Greek from Ephesians 2:10.[3] So, in this manner, we are “made things” or living poems uniquely designed with a soul with a seat of memory, passions, and a desire to know God who is more than the breath in our nostrils, but rather, who created our breath—spirit of life— and our bodily means to breathe.
Similarly, my poem, “Robot-Proof Verse,” is inspired by the Turing Test, designed by the English mathematician and early computer scientist named Alan Turing. It is a test of a machine’s ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human. It involves a judge simultaneously conversing with a human and a machine without knowing which is which—and attempting to distinguish between them based on their responses. If you’ve ever been surfing the internet and asked to verify your identity by clicking on every traffic light, freeway, or stop sign in a bunch of photographs, that’s a Turing Test.
ROBOT-PROOF VERSE[4]
Proving I am not a robot, I choose every photograph
with a bridge: in a similar drill, all the traffic lights.
How does telling a traffic light apart from a tulip
say the truth about my humanity? Does it suffice
beyond a reasonable doubt to prove I am human,
a preponderance of evidence that I harbor a soul?
Amid this ragtag crew of images, these snapshots
of scenes from anywhere, I select eight to prove
I am not a robot. Is the antithesis of robot a human?
Is there anything else I would be if I were not a robot—
a rolling stone, an armadillo, or a stick of margarine
for the buttermilk cake? A psalm of stroopwafels
or a shaggy alpaca with a snort on the other side
of the door? Will telling bridges from toll roads,
traffic lights from buses and stop signs, confirm
who I am apart from those scripted, unseen bots?
On the other side of the screen, a version of me
cannot distinguish the streetlights from traffic,
the bridges from toll booths or telephone poles.
Dare I say it—is she one of the malicious codes,
automated scripts roaming our sites mercilessly?
Scraping and mining without a mind, entering
the portals of verse where I log the imagery
of quatrains, she searches for my identity.
In what city or town were you born, my dear?
And what is your mother’s maiden name?
Where were you when you had your first kiss?
And how are your soul and spirit, I pray?
In this poem, I ponder what it means not to be a robot, ostensibly a human. The burden of proof is upon the poet to demonstrate her “humanness” by clicking on all the bridges in a group of photographs–a task, apparently, a bot cannot successfully complete. Again, this poem includes the rhetorical questions I’ve asked when I’m engaging with the world of machine learning. It also tests the limits of binary thinking– “Is the antithesis of robot a human?” The poem gently satirizes AI attempts at creativity or invention which come across as nonsensical: The sibilance of the phrase, “A psalm of stroopwafels,” is musical but nonsense. Then the poem takes a nefarious turn toward socially engineered identity theft while evoking questions about identity and authenticity, presenting security questions ending with one that’s less about username and password verification, but rather, more about our identities in Christ, germane in a soteriological sense, yet virtually useless as a security question: “And how are your soul and spirit, I pray?” In this manner, this poem, like its two previous poem-sisters, explores the tension existing between the human and the machine in this age of artificial intelligence.
For a direct application of what this could all mean in an age of artificial intelligence, let’s take a few minutes to look at an actual AI tool, such as Perplexity, and what it does with poetry. Can AI create a decent poem that doesn’t read like doggerel if prompted accordingly? What does Perplexity do when I prompt it to “Write an eight-line poem on the theme of artificial intelligence imitating the style of Emily Dickinson?”[5]
A Circuit hums—where Thought should dwell
A Mind—constructed—not by Soul—
It learns—as Shadows—mimic Light—
Yet lacks the Ache—of Human Toll—
It answers—swift—as Sparrow’s Flight—
But ponders not—the Loss—nor Goal—
A Ghost—of Reason—dressed in Code—
It haunts the House—but not the Whole—
In terms of rhyme and meter, Perplexity follows the ballad form that Dickinson used. The vocabulary and diction mirror her actual poems– Circuit, Soul, Light, Ghost, Sparrow, Reason, House, Silence, Meadow certainly appear in Dickinson’s poems, although deployed by Perplexity in a somewhat wooden manner– the staccato dashes, for instance, disrupt the poem at regular intervals. If I may, the Perplexity-generated poem feels overall tedious and mechanical. When further prompted, “How does AI mimic Emily Dickinson’s concise and enigmatic style,” Perplexity does not trot out assertions such as, “I’m just as good as Emily Dickinson,” or “I sound more like Emily Dickinson than Emily Dickinson herself” in the manner of Karaoke singers who “sound even more like or even better” than power ballad pop stars like Celine Dion, Mariah Carey, or Whitney Houston—much to my relief as a human poet.
Rather, Perplexity concedes that the original Emily Dickinson is more idiosyncratic and less predictable than the AI generated version, which relies on recognizable stylistic elements. As explained by Perplexity, its imitative Dickinsonian poems are influenced by a more formulaic bent, and ultimately “struggles to authentically reproduce her most distinctive stylistic quirks and the deeper unpredictability of her poetic voice.”[6]
I explored another artificial intelligence tool, “AI Emily Dickinson” by DeepAI.[7] I input the prompt, “Emily, what is the role of a poet or artist in this age of Artificial Intelligence?” What follows is the multi-paragraph response, which AI Emily Dickinson further summarizes as follows: “The poet’s task, in this age of machines, is to tend the flame of imagination, to nurture the eccentricities of the human spirit, and to craft a language that is both a reflection of our shared humanity and a testament to the uniqueness of each individual’s experience. The shadows that shroud the grave are, to me, an invitation to ponder the great mysteries that lie beyond the reach of human understanding. And it is in this realm of the unknown, where the machines of Artificial Intelligence may falter, that the poet’s artistry comes alive.” The AI response is a little too neat, too pat in a pre-deterministic sort of way—indeed, bearing the algorithmic imprint of generative AI in this age of automation; it falls short, if not a little dull or empty.
How do we respond to AI Emily Dickinson? At this juncture, I gently reminded the students during my visits—creative, sentient, metacognitive people designed and programmed AI Emily Dickinson, so it only knows what we’ve trained it to produce, thanks to large language models trained on enormous amounts of data to generate as sophisticated and nuanced responses as possible by analyzing data patterns. This is all to say, take courage, dear human tribe of poets and readers—existentially speaking, AI Emily Dickinson is not truly the Emily who lived, breathed, and walked this earth—neither her clone, nor her replicant, nor a twin—although we might glean insights from the products of generative AI and machine learning, there is not a unique, embodied experience of sentience for AI Emily Dickinson, as each one of us is a unique human instance in the eyes of God the Creator.
The Bulgarian-born, American essayist Maria Popova writes, “AI may never be able to write a great poem—a truly original poem—because a poem is made not of language but of experience, and the defining aspect of human experience is the constant collision between our wishes and reality, the sharp violation of our expectations, the demolition of our plans. We call this suffering.”[8] During my class visit at Wheaton, we discussed how this trait of human suffering marked Emily Dickinson’s own life, too, yielding her marvelous poems of joy, beauty, and wonder.
The poetry professor Joshua Bennett has reflected on why his students at the Massachusetts Institute of Technology are writing so much poetry—that they formed an additional poetry group beyond his class workshops. In his essay published in The Atlantic, “Why So Many MIT Students Are Writing Poetry,” Bennett writes, “The poet W.S. Merwin once said that you know you are writing a poem when a ‘sequence of words starts giving off what you might describe as a kind of electric charge.’ I’ve been thinking about how to place the sort of liveness Merwin describes—the sense of your body as a living circuit that the poem moves through—in a world filling up with noise, marred by misdirection and distraction. When, how, and why do we make room for the miraculous? From moment to moment, in any way we can, because it is part of the practice of being human.”[9]
So, dear readers and lovers of poetry, fear not: Rather, let’s conclude with a reading of a well-known piece by the original Emily Dickinson, a poem affirming the value of mystery, surprise, discovery, and the joy of slow revelation through surprising uses of language. What does our original Emily Dickinson say about truth as it relates to our poetic enterprise, how does the poem say what it means, and why is this manner of truth-telling through oblique, even elliptical, poetic compression both powerful and compelling for today’s readers?[10]
Brave poets in this age of artificial intelligence, let us be encouraged, dazzled, and delighted by the slow revelation of God’s truth through His gift of poems. As Ephesians 2:10 tells us that we are His living poems made to love God and one another, we should embrace the difficult but good labor of holistic living, rigorous thinking, and wise dreaming in the farthest reaches of the imagination where the Holy Spirit animates our hearts, and where God inhabits our praises and edifies us as image-bearers of Christ: Even in this technologically advanced age where knowledge has greatly expanded, we are still little creators made in the image of the Creator.
[1] Lucas Encarnacion-Rivera, “Semantic vs. Instance Segmentation,” Biodock, July 27, 2022. https://blog.biodock.ai/semantic-vs-instance-segmentation/.
[2] “AI Image Segmentation: How It Works (and Why It’s Important),” Cloudinary. Accessed August 7,
2025. https://cloudinary.com/guides/ai/ai-image-segmentation.
[3] John Poch, God’s Poems: The Beauty of Poetry and the Christian Imagination (St. Augustine’s Press, 2022).
[4] Karen An-hwei Lee,“Robot-Proof Verse,” Poetry Northwest 18, no. 1 (2023): 46.
[5] Perplexity. Perplexity AI. Accessed August 2, 2025. https://www.perplexity.ai.
[6] Perplexity AI.
[7] DeepAI. “Emily Dickinson – AI Chat.” DeepAI. June 24, 2025. https://deepai.org/chat/emily-dickinson.
[8] Maria Popova, “Uncoding Creativity in the Age of AI: What Makes a Great Poem, What Makes a Great Storyteller, and What Makes Us Human,” The Marginalian, July 7, 2025. https://www.themarginalian.org/2025/07/07/suffering-creativity-canetti-rilke/.
[9] Joshua Bennett, “Why So Many MIT Students Are Writing Poetry,” The Atlantic, August 15,
2025. https://www.theatlantic.com/books/archive/2025/08/what-mit-students-are-learning-
poetry/683856/.
[10] Emily Dickinson, “Tell All the Truth but Tell It Slant” [Poem no. 1263]. Poetry Foundation. Accessed
August 1, 2025. https://www.poetryfoundation.org/poems/56824/tell-all-the-truth-but-tell-it-
slant-1263.
