In November 1969, a Canadian band that I suspect almost nobody reading this will recognize—called “The Original Caste”—released a song called One Tin Soldier. I first heard the song as a child, after which its message remained with me mostly as a “love thy neighbor” style of lesson.
Some may recognize the song’s story of the mountain kingdom that possessed an untold treasure that the people of the valley below so desperately wanted to claim as their own. The folks in the valley sent a message up the mountain asking for the buried treasure, to which the people of the mountain generously replied that they’d happily share with their brothers in the valley.
I have, in more recent times, come to think of the song as a parable for our modern high-tech culture, and for artificial intelligence specifically.
You see, enraged by the mountain kingdom’s offer to share rather than to hand over their entire treasure, the valley folk drew their swords and killed the mountain people. The song’s final line:
Then they stood beside the treasure, on the mountain, dark and red. They turned the stone and looked beneath it. “Peace on Earth” was all it said.
Sometime in 2024 my father asked if I would write a new book. The purpose, he suggested, was to make the case to him and to those like him—which is to say, those who aren’t steeped in the workings of artificial intelligence—that AI is not something we should fear. Don’t Be Afraid, he thought I should title it, at least in part as an answer to his own widely shared fears.
I told him plainly that I couldn’t write it. “I am committed to the promise, to a positive vision for AI,” I said. “But I can’t tell you not to be afraid, because there are times when I am deathly afraid of what may come.”
Thus began several years writing letters to my father. About AI. Yes, my aging father and I write to one another regularly about artificial intelligence.
An odd thing happened during my lifetime. I am so ingrained in it that it took those letters for me to find the clarity to explain it.
At some point in recent memory, we—society—became fixated on features. On data. On quantifiable facts. Things like productivity and compute capacity. Never before had so many people spoken so fluently concerning technical specifics about which they really knew very little.
“Don’t lie to me,” I said during my keynote at the AI Trust Forum in London last month: “You have no idea what a megapixel is. You just vaguely suspect that your current phone has more of them than your last one.”
Michael Bloomberg, the former mayor of New York City, is known to say (though he did not originate the quote) that “In God we trust. Everyone else, bring data.”
It sounds great: Let’s make decisions based on evidence rather than emotion.
The professional class came to believe that this was clever. I certainly subscribed to the idea, and in many cases, still do. Yet…
So obsessed have we become with data, with features, with the quantifiable facts of our own supposed treasures—whatever the thing may be—that we have lost the ability to tell stories.
The valley folk sought the treasure, the quantifiable thing. They killed for it, entirely unaware that it was a story they’d find beneath the stone.
It was stories that sustained humankind’s transmission of knowledge for millennia before a horde of smug, self-impressed IT professionals and management consultants turned up in the opening decades of the twenty-first century extolling the virtues of “data-driven decisions” to an adoring world.
You might say that there’s no mythos in megapixels.
And you’d be correct.
“But my phone has forty-eight megapixels!” I can hear some past version of myself argue.
You see, almost everyone working in AI today is thinking about it as technology—features, what can it do?—or as governance—how do we control it? What policies shall we make? What laws shall we pass? Which bad behaviors or irresponsible use shall we declare unacceptable?
The engineers, psychologists, lawyers, financiers, lawmakers, marketers, managers, butchers, bakers, and candlestick makers working in AI today are almost all interested in one of three things:
How do we build it?
How do we exploit it?
How do we control it?
But I rarely hear the question asked:
Who do we wish to be? What kind of society do we wish to build with this fabulous technology?
Features and data are about building it, exploiting it. Governance is about controlling it. But only stories can answer the deeper question of who we wish to be.
If we don’t know what we want to be, then we have no means to decide which AI—or any technology—is good for us.
We must know.
Only then can we get to work bending AI towards our aims, using it to help us craft the future in our story.
Those who have read my work the last several years will recognize, here, the words of President Kennedy:
“For space science, like nuclear science and all technology, has no conscience of its own. Whether it will become a force for good or ill depends on man.”
I traveled to Athens, Greece earlier this year, looking for answers to this question in one of the cradles of human civilization. There I hired a driver to take me down the coast. I think he thought I was crazy, paying him to drive one guy ninety minutes each way so that I could walk atop a windy cliff for an hour. I think the Greeks are used to this from the rest of us. Roaming all over their creation. Gawking at piles of stones.
“It’s not always good,” the driver said to me. “Until just now, the Greeks have lived off the backs of our ancestors, selling tickets to things built more than two thousand years ago, not really building things new. But…” he pondered whimsically, “maybe now, that changes.”
I hopped out of the car. There was one odd escarpment of white rock. Below it, the translucent, sparkling blue and green of the churning sea. Beyond it on the cliffside opposite the azure inlet below, decrepit columns, an ancient monument staring out into the waters and the clouds and the setting sun beyond.
I stood in the shadow of the Temple of Poseidon, first constructed over two millennia ago in homage to the god of the sea.
The cradle of western civilization seemed as good a place as any to ponder the ethics of artificial intelligence. The sea had always felt such a part of me, or I of it. The wind and the sea and stars looking down upon it are as much a part of me as any brother or sister or ancestor, since my father first took me down to the shore and taught me to drive a boat and we stood together for a photo with my first fish.
I am struck by how much of our world was built by humans over thousands of years, each generation passing something to the next, there being so much depth of legacy in the sweep of time. And having all that knowledge consumed and understood—at least mathematically—by a machine that knows more than anyone, any human, already.
When I learned of the temple’s existence, I thought to myself… “I have to go”. I just had to go. Poseidon was the god of the sea, and hearing that he had a temple interested me far more than anything else there.
I was drawn to that temple as if it had been programmed into my basic instructions, what in the world of AI we call “system instructions”, the guidance we give to AI that tells it how to act and, importantly, how not to act. As if standing before it in the presence of thousands of years’ fellow souls, a child of the sea, was my highest, most immutable aspiration. As if people thousands of years ago had built something to honor the gods, but also to honor me. For an instant, I felt carried away by the tides of history, there, standing before a monument to the long blue line, to everything there is, was, or ever will be upon the sea.
If AI does anything—I thought, standing there—it should be freeing us to hold sacred our own humanity, our own quest for knowledge and progress, and the roots planted over generations that make a place our home.
Which returns me to the earlier question:
Who do we wish to be? What kind of society do we wish to build with this fabulous technology?
This is not a question of governance, responsibility, or even trust. It is a question of aspiration. Only in understanding our aspirations can we hope for the wisdom to use this thing that we have created.
I could not tell my father to not be afraid, but I can say:
AI must promote human flourishing.
AI must enable human exploration.
AI must preserve human culture.
Human flourishing requires that gains in productivity and wealth be paired with gains in quality of life and the freedom of the mind.
Human exploration is about putting AI to work as we cultivate our knowledge of the physical and natural world, of space and sea, and of the human body itself.
Preservation of human culture means making culture accessible to those who may not stumble upon it, refusing to diminish it in the name of progress, and putting the machine to work keeping what we might otherwise lose.
This is the work ahead. Turn the stone, and look beneath it.
Telling your dad you couldn't write Don't Be Afraid because sometimes you're deathly afraid too is the answer I'd trust. It is very easy to become pessimistic when you work with AI a lot, and I stay optimistic with my eyes open. Who do we wish to be is a much better question than how many megapixels lol