You Don’t Have to Use AI

With the rise of AI, most people feel compelled to ask: How should I be using it? But this question assumes too much. We should instead be asking: Am I free not to?

The answer is simple: Yes, you are. If you would rather not use AI — if you don’t like it, if you don’t find it useful, if you just aren’t interested in adding it to your digital toolkit — then you are allowed to opt out. There is no imperative to use it. Whatever the eventual costs, it remains perfectly reasonable to hold off for now, perhaps forever.

If this sounds like heresy, that’s because it is. The line we’re being fed, week by week, day by day, is that generative artificial intelligence is comprehensive, transformative, and irresistible. It will touch every aspect of our lives, reshaping them from the inside out. Best, then, to hop on board the train now instead of waiting until it’s too late. God forbid you be left behind.

This type of narrative, sometimes sincere and sometimes propaganda, is perennial. It applies to politics as much as to technology. It says that, because history’s shape is determined by structural changes from the top down, we can know the future before it arrives. For this reason, technologists speak in the future tense. They don’t say “might” or “maybe” or “could,” only “will.” After all, they have visited the future bearing down on us and returned, like prophets, to declare its inevitable arrival.

The paradox, however, is that if the future is inevitable, then there’s nothing for you or me to do about it. Writing in the 1940s, C. S. Lewis observed a similar dynamic in communist writers. He noted that “they tend, when all else fails, to tell me that I ought to forward the revolution because ‘it is bound to come.’ One dissuaded me from my own position on the shockingly irrelevant ground that if I continued to hold it I should, in good time, be ‘mown down’ — argued, as a cancer might argue if it could talk, that he must be right because he could kill me.”

The true believers in the AI gospel make a mistake eerily similar to that of the mid-century communists: Because they suppose themselves on the right side of history — because they imagine history has sides at all — they cannot abide even mild dissent.

But you remain free to say no.

Take Yascha Mounk, a political scientist at Johns Hopkins, who wrote an essay last summer about what he called “the peculiar persistence of the AI denialists.” He offers the “extreme” example of Jia Tolentino, a writer for the New Yorker, who “has the honesty to acknowledge the astonishing fact” that she has never used ChatGPT. Mounk ridicules Tolentino for adopting an AI-critical attitude that “resembles that of a reactionary 19th century priest who denounces the railways as the devil’s work — before proudly mentioning that he himself has, of course, never engaged in the sin of riding one.”

But here’s the thing: There really were substantial tradeoffs in the proliferation of railroads. It would not have been irrational to point this out at the time, as figures like John Ruskin did. Nor would it have been silly to criticize the railroads’ impact without having ridden a train. You don’t have to drive a car to see the exhaust pouring out of it; you don’t have to “talk” to a chatbot to know where it comes from, how it works, or what its network effects on human life are likely to be.

What’s astonishing isn’t that Tolentino is criticizing the impact of AI without having used it; it’s that Mounk believes the revolution is “inevitable,” as he puts it, and that to think otherwise is “to bury your head in the sand.”

The upshot seems to be that, if you want to be on the right side of the revolution, you should be using AI. But shouldn’t the logic run precisely the other way: that the more revolutionary — the more disruptive — a new technology seems to be, the slower we should be to adopt it? Are revolutions always simply to be gotten on board with?

Consider another example. On a recent episode of his podcast Plain English, Derek Thompson interviewed Jim VandeHei, co-founder of Politico and Axios. At one point, Thompson refers to VandeHei as “extremely AI-pilled.” So he invites him to address the skeptics.

VandeHei’s response is illuminating. He begins by saying “there’s nothing I feel more passionate about.” He revs up like an evangelist: “You’ll start to see the magic. You’ll start to see why people aren’t sleeping in San Francisco, … why people like you and I are up late at night building things.” His example: “In the last month, especially with this explosion of Claude, I’ve built a half dozen different prototype apps that are high-functioning and work.” As the adjectives pile up — magical, intuitive, beautiful — the lesson for the unconverted becomes clear: get building.

Thompson, rather than acting as the questioning journalist, doubles down on VandeHei’s revival preaching: “Pay the twenty dollars. I think Jim is right. Pay the twenty dollars and see what you can build.” The skeptics who tried an early version of ChatGPT and were underwhelmed are, he says, like people “trying out a cell phone in 1989 and then deciding that the smartphone of 2026 doesn’t work.”

This is another wonderfully revealing analogy. Like the priest on Mounk’s railroad, the 1980s cell phone skeptic is not obviously wrong, at least not at a spiritual level. Yes, the smartphone of 2026 works all too well. But if by opting out back then the cell phone skeptic could have avoided the eventual domination of the little black mirrors in everyone’s pockets and purses, she would be looking pretty good right now, even prophetic. In the words of Nicholas Carr, she might have kept her “old brain.”

To be clear, the cell phone skeptic would indeed have missed out. In certain respects, she would have been left behind. There are real tradeoffs to living without a smartphone today.

But for some of us, it’s worth it. Resistance is not for everyone, but it remains a perfectly reasonable choice. When Thompson compares early AI to early cell phones, what that says to me is: stay away.

What about those who are all in? In January, Molly Cantillon, founder of the AI assistant company NOX, wrote an essay called “AI Took Control of My Life and I Love It.” She handed over her calendar, her finances, her subscriptions, her sleep habits, “every bit of data about my life” to Claude, who, in her words, “promised salvation.” The AI agents she built to “accomplish all the tasks I needed to do” became what she herself calls “the watchtower” — an unsleeping digital panopticon. She doesn’t regret it: “The tower gave me my life back.”

Maybe. But what is the character of the life AI promises to save in this way? Cantillon was, by her own admission, overwhelmed by a schedule that wrecked her sleep and had her “tracking 15 tabs at 6 a.m., … trying to parse what the best trade is.” Perhaps the problem with her situation was not that she lacked an all-seeing intelligence capable of rearranging her life for optimized performance. It was that she felt the need to optimize every moment of her life at all.

“The sole cause of man’s unhappiness is that he does not know how to stay quietly in his room,” Pascal tells us. Many people cannot bear to imagine sitting still while the world changes around them. Even heralds of the AI apocalypse join fellow believers in the race to build artificial general intelligence.

This is the end of history and Tyler Cowen is the last man: everyone is an entrepreneur and everyone is vibe coding. The futurists stand at the terminus and with all sincerity beckon the rest of us to join them. They’re puzzled to see that not everyone is rushing forward. The only opponent they can conjure is a knuckle-dragging Luddite, uninformed by definition and perpetually behind the times.

What the true believers cannot grasp is a quotidian skeptic, an actual AI atheist, neither hysterical nor stupid. And yet a growing number of thoughtful people are looking for less from technology, not more; are deleting apps, not downloading new ones. They are doing all they can to lift their eyes from all-consuming devices to look at faces, bodies, trees, birds, sky, and sea. They are moving in the opposite direction from the believers: not into the screens but out of them.

I am not questioning whether anybody is justified in using AI. People are free to do so and many will. All the energy, pressure, and momentum — and money — in our culture are pushing in this direction.

The more interesting question is whether anybody is justified in not using it. A surprising number of people seem to me to be using AI not because they need to or want to but out of a vague sense of social or even moral obligation. They feel unable to say: No, thank you, I will wait and see whether I want to use these tools. They feel they lack agency. And so when the new tools arrive on their doorstep they sigh, shrug, and start learning how to query.

But technological change has so often had the opposite effect from what its creators promised in recent decades. Instead of empowering us, our devices and platforms have sapped and enervated us, all too quickly adding one more task on an endless checklist — Microsoft’s “infinite workday” — rather than offering a helpful means of saving time and labor.

What we need is permission: the reminder that we are free to do as we please, free to say yes or no based on our own considered judgment. We do not have to resign ourselves, heads bowed in submission. The truth is that there is no command from on high. There is no predetermined future to which technologists have privileged access. And even if there were, there would be no reason for the rest of us to help usher it in. The march of AI progress might well be worth resisting or slowing down.

There is a figure in American literature who embodies this form of resistance, this inner freedom of indomitable resolve. His name is Bartleby. He comes to us from Herman Melville’s 1853 novella “Bartleby, the Scrivener.” When given instructions by his boss, Bartleby’s reply is dispassionate but immovable: “I would prefer not to.” Over and over again Bartleby’s increasingly incredulous superior issues imperatives, and over and over again Bartleby’s answer is the same.

And when he is asked, “Why do you refuse?” Bartleby responds once more: “I would prefer not to.”

I submit Bartleby as a model to anyone who doesn’t want to embrace a bloodless AI future. Call it “The Bartleby Option.” If you don’t want to buy the latest widget, to shell out for the newest invention, to download the hottest app, you simply do not have to, and you owe no one an explanation. All you have to do is make Bartleby’s words your own and repeat them like a mantra.

Say, “I would prefer not to.” Then without another word, get on with your day, and forget the stupefied voices decreeing otherwise.