AI in Marketing

I use AI every day, and somewhere in the past year I started trusting my own thinking less. Not because the tools are bad. Because they're good. Fast, thorough, tireless. When I write something and it feels uncertain, I check it against the machine, and the machine is usually more polished than I am, which turns out to be a problem. The cognitive muscle you don't use atrophies. I'm watching mine.

I'm a marketing consultant, and AI is now central to everything I do. I use it with clients, I teach it, I get paid to find new ways to apply it. I'm not ambivalent about it. But the more time I spend inside that world, the more clearly I see what it can't do. Not for lack of processing power. Structurally. Take Dostoevsky's Crime and Punishment. A student convinces himself through careful logic that murdering a corrupt pawnbroker is morally justifiable. The argument is almost airtight. The novel then spends 600 pages on what that reasoning costs him from the inside. An AI can summarize that plot accurately. What it cannot do is put you inside the mind as the logic assembles itself, let you feel how clean it seems and how wrong it is at the same time. That's what literature does. It doesn't describe moral failure from a distance. It makes you inhabit it, and inhabiting something is a different kind of knowing than processing it. I built this list with Claude, which I realize is a strange thing to admit here. I asked the AI to help me find the books that best capture what makes us irreducibly human. It had good answers. Whether that proves the list works or quietly undermines the premise, I'll leave to you.

Twenty books, two years, one every five or six weeks. Different centuries, different continents, radically different assumptions about what a life is for. Some will resist you before they open up. That resistance is not incidental. It's the point.

Here is the list.

THE HUMAN CANON. Twenty Books for the Age of AI. A Two-Year Curriculum.

AI is very good at pattern completion, synthesis, and retrieval. What it cannot do: sit with ambiguity without resolving it, feel the weight of a moral failure, build judgment from lived contradiction, or understand why a sentence is beautiful. This curriculum is built around those capacities. Read in sequence. One book every five to six weeks. By the end you will have inhabited twenty different ways of being human. That is what AI cannot train itself on.

---

**Part One: Philosophy That Breaks Your Frame.**

We start with philosophy, but not the kind that stays on the page. These four books are destabilizing by design. They don't give you a system to adopt. They pull apart the assumptions you didn't know you were making. That's where any honest reckoning with the AI era has to begin, not with the technology, but with the question of what you actually think and why.

Book one. Meditations, by Marcus Aurelius.

AI optimizes for outputs. Aurelius optimized for character. The gap between those two projects is where human leadership lives. This is a Roman emperor's private journal, never meant for publication, written on military campaigns. It is the most honest account we have of a powerful person trying not to be corrupted by power. Short entries, high density.

Book two. The Analects, by Confucius.

Every AI system is built around individual user optimization. Confucius starts from a different premise: the self only exists in relation. That is the correction the Western tech world most needs right now. The Analects were compiled by his followers after his death. They cover ethics, governance, family, and learning through short dialogues. Read slowly. Not straight through.

Book three. The Stranger, by Albert Camus.

Meursault feels nothing the world says he should feel. In an AI world that increasingly tells you what to feel and when, his refusal is a training exercise in independent consciousness. A French Algerian man kills someone, goes to trial, and discovers he is being judged not for the act but for his emotional responses. It runs about three and a half hours. Read it twice.

Book four. Thus Spoke Zarathustra, by Friedrich Nietzsche.

Nietzsche asked: what do you do when the structures that gave life meaning collapse? That is the question of the AI era. He didn't answer it, but he asked it better than anyone. A prophet descends from a mountain to tell humanity its inherited values are finished and it must build new ones. It is deliberately difficult. You are supposed to wrestle with it.

---

**Part Two: The Moral Weight of Being Human.**

The second section moves from philosophy into something harder to name. These are books about conscience, suffering, and what it costs to act in the world. AI can process descriptions of moral failure. What it cannot do is feel the difference between knowing something is wrong and doing it anyway. These four books live in that gap.

Book five. The Book of Job, read as literature.

AI systems always try to explain outcomes. Job's wisdom is that some things don't explain, and demanding that they do is its own form of arrogance. The capacity to sit with the unexplained is uniquely human. A righteous man loses everything. His friends offer theological explanations. God appears and says nothing useful. This is the oldest serious philosophical text in the Western tradition.

Book six. Antigone, by Sophocles.

AI cannot adjudicate between two legitimate obligations that conflict. That is precisely what leadership requires every day. This is the structure of every hard decision you will ever make. A woman buries her brother in defiance of the king's law. Neither she nor the king is simply wrong. The tragedy follows from two people who are both right.

Book seven. Crime and Punishment, by Dostoevsky.

Raskolnikov uses pure logic to justify an act that destroys him from the inside out. It is the best novel ever written about what happens when intelligence runs ahead of conscience, and it is directly relevant to anyone building AI systems. A student murders a pawnbroker, convinced the logic was sound. The rest of the novel is the cost of that reasoning.

Book eight. Narrative of the Life of Frederick Douglass.

Douglass was treated as a tool. His labor extracted, his person erased. In a world where AI systems increasingly instrumentalize human attention and behavior, his insistence on the full moral weight of individual experience is the counter-argument. Written in 1845, seven years after his escape from slavery. Precise, furious, and rhetorically brilliant.

---

**Part Three: Power, Institutions, and How They Fail.**

By now you've spent time inside individual minds under pressure. This section zooms out. It looks at systems, institutions, and the specific ways that structures built by humans to serve humans end up doing the opposite. None of these books are pessimistic, exactly. But none of them are naive either. They are clear-eyed about what power does when left unexamined, which is exactly what we need to be right now.

Book nine. 1984, by George Orwell.

Newspeak, the reduction of language to limit thought, is the best existing description of what happens when content systems optimize for engagement over truth. The surveillance infrastructure Orwell imagined is now operational. Winston Smith works for the Ministry of Truth, rewriting history. He falls in love and tries to rebel. When you get to it, read the Appendix on Newspeak.

Book ten. The Peloponnesian War, by Thucydides.

There is a section called the Melian Dialogue, five pages in which Athens tells a weaker city that might makes right. Those five pages explain more about how technology companies interact with regulators and smaller competitors than most business books combined. Athens and Sparta destroy each other over 27 years. Thucydides was there for parts of it. The Melian Dialogue alone is worth your time.

Book eleven. One Flew Over the Cuckoo's Nest, by Ken Kesey.

The institution in this novel doesn't hate its patients. It processes them efficiently for compliance. That is the risk of any system, including AI systems, that optimizes for order over humanity. McMurphy enters a psychiatric institution and refuses to be managed. Nurse Ratched is not evil. She is a perfect bureaucrat. The horror is in the gap between those two things.

Book twelve. Long Walk to Freedom, by Nelson Mandela.

Mandela maintained strategic clarity, moral purpose, and the capacity to forgive, simultaneously, over 27 years of imprisonment. That combination is the highest form of human leadership. No algorithm produces it. This is the autobiography of a man who emerged from prison to lead his country without revenge. Not an inspiration text. A case study in sustained judgment under conditions most people cannot imagine.

---

**Part Four: Empathy Across Difference.**

We are now at the midpoint, and it's worth pausing on what the list has done so far. Rome, ancient China, colonial Algeria, nineteenth-century Russia. We've been moving across time and geography deliberately, because one of the things AI gets wrong at scale is assuming that one kind of experience, usually Western, usually recent, usually English-language, is the default. The next three books push directly against that assumption. They ask you to inhabit a perspective that the dominant culture either ignored or actively tried to erase.

Book thirteen. Things Fall Apart, by Chinua Achebe.

The novel demonstrates what happens when one worldview, certain of its own rationality, encounters another and cannot recognize it as coherent. AI training data encodes exactly this kind of asymmetry at scale. An Igbo warrior in 1890s Nigeria watches his civilization be dismantled by missionaries and colonial administrators who cannot see what they are destroying. Read this before the European novels, not after.

Book fourteen. Season of Migration to the North, by Tayeb Salih.

A Sudanese man returns home after years in Europe, carrying both worlds inside him without belonging fully to either. The experience of cultural in-betweenness gives certain people an irreplaceable perspective that AI cannot replicate. This is the Sudanese answer to Heart of Darkness, written from inside the postcolonial experience. Short, dense, and morally unresolved by design. One of the great Arabic-language novels.

Book fifteen. Waiting for Godot, by Samuel Beckett.

Two people wait for something that never arrives and fill the time with language that circles without landing. In a world of infinite content, Godot asks: what are you actually waiting for, and what do you owe the person next to you while you wait? Written after the Holocaust. Two men, a road, a tree, a boy who comes and goes. Nothing resolves. The point is what they do anyway.

---

**Part Five: Knowledge, Bias, and How We See.**

Everything we've read so far has been, in some sense, about experience. This section is about knowledge itself. How it gets made, who gets to make it, and what happens when the process is less neutral than it appears. These are not abstract questions. Every AI system is trained on data that someone selected, labeled, and decided was representative. Understanding how knowledge encodes power is not an academic exercise anymore. It's a survival skill.

Book sixteen. Orientalism, by Edward Said.

Said demonstrated that what looks like objective knowledge is often the perspective of the powerful, formalized into scholarship. AI training data is the most consequential version of this problem ever created. This book is not optional for anyone working in AI. Published in 1978, Said argues that Western scholarship on the Middle East and Asia was not neutral description but a tool that served colonial power by defining what the so-called Orient was. The argument holds, and the method applies everywhere.

Book seventeen. Man's Search for Meaning, by Viktor Frankl.

Frankl argues that meaning is not found in the environment but constructed from within. In a world where AI will absorb more of what used to give people purpose, the capacity to generate meaning independently becomes the central human skill. A psychiatrist survives the Nazi concentration camps and builds a theory of meaning from the experience. The first section, the memoir, is devastating. The second, the theory, is where it becomes a working tool.

Book eighteen. Frankenstein, by Mary Shelley.

The novel's actual question is not what did the creature do. It is what does the creator owe the created. Every person building AI systems should read this as a direct brief on their responsibilities. Victor Frankenstein creates life and abandons it. The creature's response is not madness. It is a logical consequence of rejection. The monster gives one of the most coherent moral arguments in English literature.

---

**Part Six: Cultural Breadth and the Long View.**

We end where we should: with the long view. These two books are separated by centuries and continents, but they share something. Both are about what happens after the ambition, after the striving, after you've optimized everything that can be optimized and still feel the question underneath it all. In a curriculum about staying human in the age of AI, this felt like the right place to land.

Book nineteen. Siddhartha, by Hermann Hesse.

The novel is about what happens when you achieve every goal you set and discover it wasn't the point. This is the trajectory of many careers in the technology industry. The earlier you read this, the more useful it is. A young man in ancient India leaves his privileged life, seeks enlightenment through asceticism, fails, tries pleasure and wealth, fails again, and eventually finds something quieter and harder to name. Short. Read it at a moment of transition.

Book twenty. Don Quixote, by Cervantes.

This is the first novel in the Western tradition that knows it is a novel. A man who has read so many stories that he can no longer tell them from reality. In a world where AI generates infinite narrative at scale, Cervantes already mapped the territory. An aging Spaniard reads too many chivalric romances and sets out to be a knight. His companion Sancho Panza is sane. Together they cover more ground about fiction, reality, loyalty, and delusion than most philosophy. It is long. Listen across several months.

---

**The Reading Order.**

Start here, in this sequence. One book every five to six weeks.

One: Meditations, Marcus Aurelius. Two: The Analects, Confucius. Three: Antigone, Sophocles. Four: The Book of Job. Five: The Stranger, Camus. Six: Things Fall Apart, Achebe. Seven: Narrative of the Life of Frederick Douglass. Eight: Crime and Punishment, Dostoevsky. Nine: Season of Migration to the North, Tayeb Salih. Ten: Orientalism, Said. Eleven: 1984, Orwell. Twelve: The Peloponnesian War, Thucydides. Thirteen: Frankenstein, Shelley. Fourteen: Waiting for Godot, Beckett. Fifteen: Thus Spoke Zarathustra, Nietzsche. Sixteen: Siddhartha, Hesse. Seventeen: Long Walk to Freedom, Mandela. Eighteen: One Flew Over the Cuckoo's Nest, Kesey. Nineteen: Man's Search for Meaning, Frankl. Twenty: Don Quixote, Cervantes.

Two years. Twenty ways of being human. That is what AI cannot train itself on.