The AI Odyssey: How a primal worry from the Ancient Greek world has been woke up by new expertise

Odysseus ended up lashing himself to the mast and ordering his crew to plug their ears with wax. When he heard the sirens’ call, though, he screamed to be released and begged his men to let him go. The man who’d spent 10 years outwitting gods didn’t trust himself for 30 seconds within earshot of a sound he knew would destroy him.

That specific kind of dread—the fear of what you might do if the temptation were real—is older than the internet, older than AI, older than any technology that has ever been accused of making us weaker. It was a key plot line in the Iliad and The Odyssey, and centuries later, became the central problem of Greek philosophy, the thing the Stoics spent their lives trying to solve. And it is the same thing a large and quietly growing number of people feel every time they open a chatbot; why they hate the technology without using it, and want to bring to a screeching halt the data centers driving the biggest economic boom in at least a century.

It’s the fear about how you can’t help yourself, that you’re not good enough or smart enough; that when an easy, even potentially destructive solution lies in front of you, you won’t be strong enough to resist the call of the sirens. Socrates had the same fear about the act of writing, scribes had the same fear about Gutenberg’s printing press, and John Saxon had the same fear about the calculator. But AI concentrates all of these fears in a place beyond the act of writing, printing or calculating to the very act of thinking itself. Seen in this light, The Odyssey‘s billion-dollar success isn’t an accident for times like these.

The fear underneath the data

People who secretly use AI at work and describe it as “cheating” aren’t worried about their boss finding out. The word surfaces because some part of them is already asking a different question: if this were taken away tomorrow, would I still be good at my job? Would I still be smart?

Everyone knows the name for this: impostor syndrome, the persistent, private fear of being exposed as less capable than everyone believes. With AI, there’s now a tool sitting by your side, silently keeping score—and exposing every hallucination that you’re not qualified enough to catch.

The second fear is about character, not competence: what if I don’t actually have any self-control? What if I accept the first output from the chatbot without thinking of a way to genuinely improve it? What if I get roasted for putting more AI slop into the world?

The psychological literature calls this “cognitive offloading”—using external tools to reduce the internal demands of a task. Humans have done this since the days of the ancient Greeks, with Socrates one of the earliest recorded voices in history to worry about how humans let themselves off the hook of heavy intellectual lifting. Generative AI offloads reasoning, synthesis, and creative judgment—the cognitive work through which expertise is actually built, not just retrieved. Social psychologists have documented for decades how people are “cognitive misers,” a term coined by Susan Fiske and Shelley Taylor in their foundational work on social cognition, meaning we find effortful deliberate thought aversive and default to shortcuts whenever available.

This may be the real controversy over AI that is roiling the workplace and society, as noted in The AI Shift, the Financial Times‘ weekly newsletter on how AI is reshaping work by columnist John Burn-Murdoch, who drew on research from economists José Azar, Mireia Giné and Javier Sanz-Espín to show that the most AI-exposed occupations aren’t shedding workers—employment, on paper, looks fine. But wages in those roles are down sharply relative to less-exposed work, and workers in them have become far less likely to switch jobs at all. Workers aren’t getting fired as a result of AI, but they seem to be getting quietly worth less, and quietly stuck. The real fault line, Burn-Murdoch argues, isn’t between occupations, but between execution and evaluation.

Burn-Murdoch’s name for what separates these two groups is conscientiousness—and he notes that large numbers of workers who use AI on the job do so secretly, describing it as “cheating.” But in the summer of Odysseus, workers in an AI-addled word seem to fear something more primal and ancient: that they can’t resist the sirens’ call.

Every technology that saves you effort has done this before

Back to Socrates, for a second: the great Greek philosopher made eerily familiar arguments in opposition to the creation of the technology of writing, around 370 BCE. He argued that writing would “implant forgetfulness,” producing students with the mere appearance of wisdom—people who can recite but can no longer actually think. Scrolls would turn humans into stochastic parrots, he said, to paraphrase a modern critique of AI.

The same fear resurfaced, smaller and less remembered, when the printing press began moving knowledge out of memory and onto the page. A German Benedictine abbot named Johannes Trithemius wrote a treatise called De Laude Scriptorum (“In Praise of Scribes”) arguing that printed books would never be the equal of handwritten manuscripts. “Printed books will never be the equivalent of handwritten codices, especially since printed books are often deficient in spelling and appearance.”

The fear resurfaced again, with real data behind it this time, when calculators entered American classrooms in the 1970s and ’80s. The controversy peaked in 1986 when math teacher John Saxon led a protest at the National Council of Teachers of Mathematics meeting. Subsequent research actually backed him up, finding that habitual calculator use before students had internalized basic arithmetic weakened their mental fluency and fact retrieval. But great writers and great mathematicians were able to overcome their inner cognitive miser and use these new technologies to great effect.

The Greeks had a word for this

The Greeks did not just casually worry about self-control. They built an entire branch of philosophy around it. They had a precise word for the exact failure mode: taking the shortcut while already knowing better. They called it akrasia—literally “without power”—acting against your own better judgment, with full awareness that you’re doing it. Socrates denied it was even possible. A generation later, Aristotle broke with him and insisted the problem was real, then built something closer to a spectrum of character to explain it. At one end is sophrosyne, or temperance: your appetites are already trained toward the right thing, so there’s no battle. Below that is enkrateia: you want the shortcut, reason wins, but you feel the pull the whole time. Below that is akrasia: desire wins. At the very bottom is akolasia—indulgence so habitual that you don’t register the shortcut as a temptation anymore. It’s just what you do.

The crucial move in Aristotle’s account is that none of these are fixed identities. You move along that spectrum through repetition. You become someone who doesn’t feel the pull of the easy path by doing the hard thing over and over until it becomes who you are. That is, with almost no translation required, the exact difference between the Chinese students who used AI like a tutor and kept practicing, and the ones who used it to skip the practice altogether. One group was quietly building sophrosyne. The other was drifting toward akolasia without noticing the drift. The Stoics went even further and made self-mastery close to the entire point of living a good life.

Christopher Nolan spent hundreds of millions of dollars and every inch of 70mm IMAX film stock this summer making the case that the Odysseus myth is exactly the one our moment needed. The audience agreed: it opened in July and blew past $1 billion at the global box office, becoming Nolan’s biggest film to date.

I myself have played a role in the debate over AI writing—one that has taken on new significance since Harvard economist Ricardo Hausmann admitted to using it in a column for the Financial Times and Stanley Druckenmiller did the same for a column in the Wall Street Journal (the same publication profiled my use of the tech back in March). Many people, even close friends, have told me that my writing sounds smart but they don’t believe it’s really me, and sometimes that hurts my feelings, but these are weird times.

What I stress is how I use the technology as a tool and push myself and it to go deeper, not settling for the first output. This piece, for instance, included prompts about Burn-Murdoch’s idea of “conscientiousness” and my questions about what broader pattern does this fit into in terms of human psychology and what previous technologies surfaced similar tensions. (Claude rejected a musing of mine that it brought to mind the movie Idiocracy, suggesting Wall-E instead, but I dismissed that angle. Maybe I shouldn’t have; this is long enough.)

But driving me on the whole time was my hunch that, deep down, all of us humans have a secret fear that we are a bit stupid and have no self-control. Socrates himself had the same fear, after all, and so did Odysseus.

For this story, Fortune journalists used generative AI as a research tool. An editor verified the accuracy of the information before publishing.