Can AI Understand To Kill a Mockingbird?

Maurício Pinheiro

Atticus Finch, My Father, and the Difference Between Reading a Story and Living One

Some books remain with us because they are great works of literature. Others remain because, somewhere along the way, they become part of our own biography.

For me, Harper Lee’s To Kill a Mockingbird (1960) is both.

I first read the novel as a child. I remember it almost intuitively: the shadow of Boo Radley, the trial of Tom Robinson, Scout’s bewilderment before the adult world, and above all, Atticus Finch. At that age, Atticus seemed self-evidently admirable: calm when others panicked, rational when others surrendered to prejudice, willing to stand alone when standing alone was the price of integrity.

Years later, reading the novel again as an adult, I discovered something obvious only in retrospect. The book had not changed. I had.

A machine can compare two copies of the same text and confirm that every word remains where it was. What it cannot obtain from the text alone is the history of the reader who returns to it decades later. That difference—the stability of the book against the movement of a life—would eventually become, for me, one of the most interesting questions raised by artificial intelligence.

By then, however, Atticus had already become inseparable in my mind from someone real: my father.

My father is a nuclear scientist and, as a teenager, spent time in Alabama as an exchange student during an extraordinary period in American history. The United States was entering the Space Age, competing furiously with the Soviet Union in physics, rocketry, and nuclear power. Yet on the ground, racial segregation still governed daily life across the South. Sputnik had transformed the night sky into a symbol of technological ambition and Cold War anxiety, while below it, an older and far more brutal injustice remained firmly entrenched.

My father encountered that racial prejudice firsthand.

There is something almost unbearable in that historical juxtaposition. A civilization could split the atom, build rockets, calculate orbital trajectories, and send machines beyond the atmosphere while remaining morally primitive in its treatment of millions of human beings. Technological sophistication had outrun moral progress.

Perhaps that is why, when I returned to Harper Lee’s Alabama years later, Atticus no longer felt like a mere literary character. He gave shape to qualities I had already known in my own father: intellectual independence, restraint, quiet courage, and the refusal to accept the verdict of a crowd simply because it was the crowd’s verdict.

I do not mean that my father was Atticus Finch, or that either man should be transformed into a moral saint. Literature becomes less interesting when its characters are reduced to monuments, and memory becomes less truthful when we do the same to people we love. My connection was simpler and more personal: Atticus gave literary form to qualities I already recognized.

He did not teach me what they looked like.

He helped me see them more clearly.

The Mockingbird

The plot of To Kill a Mockingbird is deceptively simple. Atticus defends Tom Robinson, a Black man falsely accused of rape in 1930s Alabama. Scout watches the tragedy unfold through the unfinished understanding of childhood, and precisely because she has not yet learned all the evasions by which adults protect themselves from their own contradictions, she often sees those contradictions more clearly than they do.

The novel is not disturbing because prejudice is difficult to recognize. Prejudice is often blatant. What disturbs us is how easily ordinary people learn to live beside it.

The citizens of Maycomb invoke faith, family, law, and respectability while participating in a social order that violates those very principles. Scout gradually discovers that moral language does not necessarily produce moral behavior; people can sincerely believe themselves decent while accommodating an indecent system.

This is also where an adult reading of Atticus becomes more complicated than the admiration of childhood. His courage in defending Tom Robinson remains real, but courage inside an unjust system does not automatically redeem the system itself. Later debates surrounding Atticus—and the more troubling version of him presented in the subsequently published Go Set a Watchman (2015)—make it difficult to preserve him as an uncomplicated moral icon. That complication does not destroy the character for me. It makes the question more interesting: what can individual decency accomplish when the institutions surrounding it are profoundly unjust?

The mockingbird — the Northern Mockingbird (Mimus polyglottos), a North American bird known for imitating the songs of other birds and the species that gives the original English title its name — condenses this problem into an apparently delicate image. The word has no perfect equivalent in Portuguese: although it is sometimes rendered as cotovia or tordo-imitador, neither fully preserves the cultural and symbolic meaning that mockingbird carries in the novel.

It harms no one; it simply sings. Its destruction becomes a metaphor for a recurring human habit: turning vulnerability into guilt. Societies under pressure often need someone onto whom they can project fear, resentment, or a threatened sense of identity. The victim changes; the mechanism persists with remarkable consistency.Tom Robinson is the clearest victim of that mechanism. Boo Radley, in another way, is one too.

The metaphor survives because the need for scapegoats survives.

Different eras merely choose different mockingbirds.

Recognition Without Remembrance

Decades after my first encounter with Maycomb, I found myself doing something that would have sounded like science fiction when I was a boy: discussing Harper Lee with an artificial intelligence. I first wrote this article in January 2023, only a few weeks after ChatGPT became publicly available. At the time, the novelty of conversing with a machine about literature was still striking.

What began as a curiosity soon became a more interesting question: not whether AI could summarize a novel, but what its way of reading might reveal about our own.

That experience matters for a reason deeper than the obvious fact that a machine has no childhood.

Ask a capable language model about To Kill a Mockingbird, and it can do far more than summarize the plot. It can reconstruct the historical context, analyze Scout’s narrative perspective, trace the symbolism of the mockingbird, compare competing interpretations of Atticus, examine the legal and moral architecture of Tom Robinson’s trial, and connect the novel to broader questions of prejudice, justice, conformity, paternalism, and institutional power—usually in seconds, across a range of sources and perspectives that would take a human scholar considerably longer to assemble.

(My apologies to literature departments: the machine has not abolished literary criticism yet, but it has made charging tuition merely to summarize themes and symbolism a rather more difficult business to defend.)

Some of those interpretations can be genuinely useful. Some may expose connections a human reader missed.

So the interesting question is not whether the machine “really feels” the novel. That is too easy.

The interesting question is what the existence of such a reader reveals about our own reading.

A language model can associate Atticus Finch with moral courage, paternal authority, legal justice, racial inequality, and hundreds of related concepts. I encounter the same character and remember my father.

The difference is not that one process contains meaning and the other does not. It is that they arrive at meaning through different histories.

The machine reaches Atticus through relationships among texts, concepts, patterns, and representations. I reach him through those things too—but also through a life that existed before I opened the book again.

When I write, “Atticus Finch reminds me of my father,” an AI can analyze the comparison, identify the traits that support it, challenge its assumptions, and even help me articulate aspects of the association I had never consciously formulated.

But the sentence did not originate in literary analysis.

It originated in memory.

The machine encounters Alabama as a dense intersection of geography, segregation, legal history, literature, politics, archives, and language.

I encounter all of that—and my father.

That extra connection is not necessarily more intelligent. It is more particular.

And its particularity is precisely what makes it mine.

Intelligence Without Biography

Artificial intelligence therefore forces a question that literary criticism could once leave largely implicit:

What does it mean to understand a book?

If understanding means identifying structures, recognizing symbols, reconstructing contexts, comparing interpretations, and discovering relationships among ideas, then artificial intelligence already demonstrates a remarkable form of literary understanding.

But reading has never consisted only of extracting meaning from a text. Meaning also changes according to the life that encounters it.

We do not approach a book as empty processors. We arrive carrying a life.

That life rarely announces itself in grand philosophical categories. It appears in details: the sound of a parent’s voice, a room remembered imperfectly, a sentence someone once repeated, a street we have not seen for decades, an argument whose importance became clear only years later. A fictional house may suddenly resemble one we once entered; a minor character may acquire the mannerisms of someone we knew; a passage that seemed unremarkable at twenty can become almost impossible to read at sixty.

The words have remained fixed while the reader has moved through time, accumulating experiences that alter what those words are capable of awakening.

This is why biography matters to interpretation—not because personal memory is superior to analysis, but because it introduces contingencies that no text contains by itself.

A language model may construct extraordinarily rich relationships among words and ideas. A human reader can do that too, but those relationships are continually interrupted, distorted, enriched, or illuminated by the accidents of having lived one particular life rather than another.

That distinction becomes especially interesting because AI does not merely reveal something about machines.

It reveals a danger for us.

The Risk of Reading Like a Machine

The most consequential question may not be whether artificial intelligence can learn to read more like human beings.

It may be whether human beings will begin reading more like artificial intelligence.

We already live inside an economy of summaries, excerpts, recommendations, search results, rankings, generated explanations, and compressed information. Increasingly, we approach books through systems that tell us what they contain before we have experienced them ourselves. We can ask for the themes, symbolism, historical significance, key quotations, character arcs, and conclusions without passing through the slow uncertainty of reading the work.

There is enormous convenience in this. There is also a subtle danger.

A novel can gradually become something to be processed rather than something to be undergone.

If I ask an AI for the meaning of To Kill a Mockingbird before reading it, I may receive an excellent explanation. But an explanation delivered in advance can also narrow the range of what I am prepared to discover. The categories arrive before the experience: racism, innocence, justice, moral courage, childhood, empathy. All correct. All useful.

And yet literature does not always matter because it gives us categories.

Sometimes it matters because something happens while we are reading that no summary could have predicted.

Atticus becomes my father.

That possibility depends partly on inefficiency: on wandering through a book, misunderstanding it, returning to it years later, noticing what we once ignored, and allowing our own history to interfere with the text.

Algorithmic culture encourages the opposite instinct. It offers compression, retrieval, relevance, optimization. Those are extraordinary intellectual tools, but literature has never been entirely optimized for efficiency.

Some books need time because the reader needs time.

Alabama, Then and Now

This brings me back to the image that has haunted this essay from the beginning: Alabama beneath the sky of Sputnik.

The parallel with our own technological moment should not be overstated. The moral catastrophe of racial segregation and the epistemic limits of artificial intelligence are not equivalent problems.

But the historical juxtaposition offers a warning.

The America of the Space Age demonstrated that extraordinary technological sophistication does not automatically produce equivalent sophistication in every other dimension of human life. A society can master nuclear physics while remaining blind to obvious injustice.

Our own era is producing another astonishing technology. We are building systems capable of manipulating language, knowledge, images, argument, and perhaps eventually forms of reasoning at a scale that would have been almost unimaginable a generation ago.

It would be equally naive to assume that such technological power automatically contains wisdom, memory, judgment, moral imagination, or the depth we associate with lived experience.

But it would be just as naive to conclude that humans possess those qualities automatically merely because we are human.

History gives us too much evidence to the contrary.

The real contrast, therefore, is not between a soulless machine and a profound human being.

It is between different ways of encountering the world—and between what each can reveal about the limitations of the other.

Reading After AI

This is why I do not see artificial intelligence as an enemy of literature.

It may become one of its most interesting companions.

But that depends on what we ask it to do.

If we use AI to avoid reading, it becomes a machine for intellectual compression.

If we use it after reading—to argue with us, surprise us, expose weaknesses in our interpretation, or illuminate what we have failed to notice—it can become something more valuable: a counterpoint to our own limitations.

For centuries, there was little reason to separate intelligence from biography because every intelligence capable of discussing a novel also possessed some form of human past. That assumption no longer holds. We can now converse with systems capable of analyzing grief without having buried anyone, childhood without having been children, nostalgia without having grown old, and courage without ever having been afraid.

The point is not that this makes their analysis false.

The point is that, for the first time, analysis and biography can be visibly separated.

And once separated, we begin to see what each had been contributing all along.

When I was young, I thought Atticus Finch was one of the reasons I loved To Kill a Mockingbird. Much later, I realized that I had the direction partly wrong. Literature had given me a fictional form in which I could recognize qualities I already associated with a real man.

My father.

Artificial intelligence can now help me examine that recognition. It can challenge it, complicate my reading of Atticus, reconstruct the historical Alabama in which Harper Lee wrote, and connect my private association to questions much larger than my own family.

That is not a trivial contribution.

But the original connection was not contained in the novel, waiting to be extracted like information from a database. Nor was it waiting inside the machine.

It emerged from the encounter between a book and a life.

When Atticus Finch walks onto the page today, I can analyze him differently than I did as a child. I can see the limitations of the character, the complexity of the society around him, the history behind the novel, and the arguments that have accumulated around it.

An AI can help me see all of those things.

But before any analysis begins, something else still happens.

I remember my father.

And perhaps that is what reading after AI can teach us: not that machines are incapable of understanding stories, but that understanding has never been only a property of the story.

Part of it belongs to the life that arrives to meet it.