It’s not what AI cannot do.
AI is doing what we thought only humans could do
At Wimbledon in 2023, IBM deployed an AI announcer that doesn’t just describe tennis. It gets excited about it. When a player makes an impossible return, the AI’s voice rises. During tense rallies, it builds suspense. The system was trained on 30,000 hours of audiobooks to learn how humans express emotion, then learned tennis itself.
What was once considered beyond the reach of machines, interpreting and expressing human emotion in real time, is now deployed and commercially scaled.
The same shift is happening in design.
In 2025, Archinect ran an experiment. Five cover letters for an architect position. One written by a human, four by AI. Only 28% of readers identified the human-written one. More striking: 40% rated the AI’s letter as the best. Not the most authentic. The best.
In Atlanta’s West End, an AI-powered design firm called cove designed a 16-townhome affordable housing development. The system scanned building codes, interpreted zoning rules, ran cost simulations in real time. Results: timelines cut in half, development costs down 20%.
And it runs wider. A fully AI-generated actress, Tilly Norwood, drew talent agency interest in Hollywood. An AI band, The Velvet Sundown, accumulated over a million Spotify streams before anyone realized no human made the music. Albania appointed an AI system to oversee the country’s public procurement and combat corruption.
AI is not just automating the repetitive. It is doing things we believed required something distinctly human: emotional attunement, creative expression, professional judgment.
These stories matter beyond their economic implications. They expose a question about human essence we have long been able to defer. As long as creativity and emotional expression seemed self-evidently beyond machines, we did not need to examine what actually grounds human being.
AI is removing that deferral. The question is no longer hypothetical.
A familiar pattern
None of this is unprecedented.
Before AI, we thought only humans could calculate, until computers proved otherwise. We thought strategic thinking was uniquely human, until IBM’s Deep Blue defeated world champion Garry Kasparov at chess. Later, Google’s AlphaGo beat the world’s best Go players, mastering a game long considered too intuitive for machines.
Each time a technology crossed what we assumed was a distinctly human threshold, the same two responses followed. First: the next boundary will hold. Then, when it doesn’t: what, if anything, remains uniquely ours?
We are in that second moment now.
But the pattern itself tells us something. If the boundary keeps moving, perhaps the boundary was never the point. Perhaps what we are is not defined by where the line is drawn, but by the fact that we are the ones drawing it, the ones anxious about it, the ones who decide what to do with the capabilities we keep creating.
“If the boundary keeps moving, perhaps the boundary was never the point.”
If that is true, the question itself needs reexamining. What makes us unique may not be the right question to begin with.
Capability is not essence
The obsession with what makes us intellectually or creatively unique assumes that human value is grounded in capability. But consider: dogs hear frequencies we cannot. Cheetahs run three times faster. Eagles see eight times farther. We do not feel threatened. We do not search for what humans can do that dogs cannot and conclude that defines our worth.
A dog is a dog because of what it is, not what it can do.
So why does AI feel different?
Partly because it is our own creation, made to simulate our minds. When it produces outputs indistinguishable from ours, it functions like an uncanny mirror. We recognize ourselves in it. That recognition unsettles something deep.
But there may be a second reason, less philosophical and more honest. The question of what only humans can do is often an economic question in disguise. What it really asks is: what can I still be paid for? What is my market value when my skills can be automated?
“The question of what only humans can do is often an economic question in disguise. What it really asks is: what can I still be paid for?”
This anxiety is particularly acute for professionals whose entire identity has been built around specialized expertise. When that expertise becomes replicable, the threat feels existential. Not because we stop being human, but because our economic value is lost.
This is a legitimate concern. But it is a different concern than the one about human essence. When AI displaces professional capability, we instinctively reach for an economic framework and measure ourselves by what we can still produce that a machine cannot. That framework works for economics. It is not fit for defining our being.
Being precedes doing
The boundary between human and machine capability will continue to move. AI will eventually cross thresholds we currently believe it cannot. And when it does, the question of what makes us unique will keep retreating.
The more essential question is not what AI cannot do, but what it means to be the ones who built it, who deploy it, who live in the world it is reshaping. We have to live with the impact and consequences of what we produce.
Human essence, in this light, is not a list of capabilities that machines have yet to replicate. It is something prior to capability. Something closer to what it means to be a being for whom any of this matters at all.
Being precedes doing.
“Being precedes doing.”
The search for human essence has not been misplaced because the question is unanswerable. It has been misplaced because we have been looking in the wrong direction, toward what we can do, rather than toward what we are.
We have been reluctant to look in that direction. Capability is measurable. Being is not. But the question remains. This shift in direction does not make the question easier. It may make it harder. But it points us toward what has been there all along.
Now back to the question: what are we?
Not the creatures who can design buildings faster than machines, but the ones who have to dwell in them. Not the ones who can generate solutions, but the ones who bear the weight of consequences.
References
Barrows, K. (2025, February). “AI Wrote a Better Cover Letter Than Me, I Don’t Care.” Archinect. https://archinect.com/features/article/150494228/ai-wrote-a-better-cover-letter-than-me-i-don-t-care
cove. (2025, August). “AI-Designed Affordable Housing in Atlanta’s West End.” https://cove.inc/blog/ai-designed-affordable-housing-atlanta-west-end/
Feder, T. (2024, December). “The Future of Tennis Broadcasting with AI Sports Commentary.” IBM Think. https://www.ibm.com/think/news/future-tennis-broadcasting-ai-sports-commentary
Rankin, J. (2025, September). “Albania Appoints World’s First Virtual Minister.” Politico. https://www.politico.eu/article/albania-apppoints-worlds-first-virtual-minister-edi-rama-diella/
