Yesterday morning I watched a video about artificial intelligence, scientific discovery, and what happens when machines become capable of doing work humans have historically considered uniquely ours. Yesterday afternoon, my daughter came home from school and told us that one of her classmates has leukemia.
Those two things should probably have remained completely unrelated. In my head, they became inseparable.
The video eventually turned toward a debate about automating science. It referenced an argument made by NYU professor and AI researcher Julian Togelius after a panel at NeurIPS in 2025. Togelius was disturbed by researchers openly discussing a future in which AI might eventually replace humans at every level of scientific research. He stood up during the panel and called the idea evil.
His reasoning wasn’t stupid. In fact, it was deeply human. Scientists love doing science. Young researchers dedicate years of their lives to becoming capable of expanding human knowledge. Scientific discovery provides meaning, identity, purpose, community, and wonder. If machines eventually become better at the entire process, we could deprive generations of people of one of humanity’s most meaningful intellectual activities.
I understand that argument.
Then someone essentially asked him how far he was willing to take it. What if automating science meant curing cancer faster? His answer was startlingly clear: he did not think curing cancer faster was worth it if doing so meant humans could no longer meaningfully participate in science.
And that stopped me.
Because whatever philosophical sophistication surrounds the argument, I find that moral calculation almost impossible to accept. Not because human participation doesn’t matter. Not because scientific discovery isn’t meaningful. Not because I want human beings sitting around passively while machines inherit every interesting thing left to do. It bothers me because the priorities seem inverted.
Human participation is valuable. Human suffering is more important.
Then my daughter came home.
One of her classmates has leukemia. I don’t know the details of this child’s diagnosis, and they aren’t mine to share. I don’t want to turn someone’s sick child into a rhetorical device. But I can say what learning about her diagnosis did to me: it transformed the word cancer.
“AI might help cure cancer” is the kind of sentence we’ve all heard so many times that it barely means anything anymore. Cancer becomes shorthand for some enormously important future scientific problem, lumped together with fusion energy, climate change, longevity, food scarcity, new materials, and every other promise we attach to technological progress.
Then cancer becomes a little girl who sits in a classroom with my little girl.
Suddenly the abstraction has a face.
Somewhere in the enormous possibility space of future artificial intelligence, perhaps there is a system capable of finding something human researchers would otherwise discover ten years later. Or five years later. Or one year later. Maybe it isn’t “the cure for cancer,” because cancer isn’t one disease and reality is considerably messier than the slogans we use when talking about technology. Maybe it is a molecule, a drug combination, a diagnostic technique, a better way of matching a treatment to a particular child, or simply a research direction no human being thought to pursue.
I don’t know. Nobody does.
But if that possibility exists, then I find myself asking a question I can’t quite shake: how much additional human suffering should we accept merely so that we can remain the ones making the discovery?
That question followed me into another thought, and this is where things get uncomfortable.
Every day I open social media and see people I know raging against artificial intelligence. Artists, designers, writers, photographers, musicians. People I genuinely like. Many of their concerns are legitimate. Concentrated corporate power deserves scrutiny. Copyright deserves serious discussion. Workers should not simply be discarded while a handful of technology companies capture every economic gain. Surveillance, misinformation, environmental costs, and ownership of these systems are enormous problems. None of that should be waved away in the name of progress.
But there is another strain of resistance that bothers me much more. It is harder to state politely because, stripped down, it sounds something like this: I enjoyed doing something that a machine can now do, therefore I don’t want the machine doing it.
I liked designing flyers. I liked making restaurant menus. I liked writing advertising copy. I liked illustrating books. I spent twenty years becoming good at this. This is part of who I am. Therefore the technology threatening it must be resisted.
Sitting with the knowledge that a child my daughter knows has leukemia, the absurdity of that position became difficult for me to ignore.
Obviously, the person angry that AI can make a restaurant menu has not consciously chosen their freelance design work over a child’s cancer treatment. That would be a grotesque caricature, and it is not what I believe. In fact, many of the people I see fighting AI most passionately are deeply compassionate people. Some of them are my friends. These are people who, if a child were standing in front of them and some impossible choice required it, would probably cut off their own arm to save that child’s life.
I believe that completely.
And that is precisely why the contradiction fascinates me.
The extreme comparison is useful not because it exposes hidden cruelty, but because it exposes a hidden tradeoff. The threatened thing is immediate: my job, my income, my craft, my identity, my years of training, my sense of being useful. The possible benefit is distant and distributed: some drug discovered somewhere, some scientific problem solved faster, some child I will never meet living instead of dying, some energy breakthrough arriving years earlier, some improvement in food production occurring on another continent.
One of those things touches the ego directly. The other arrives as an abstraction.
Of course the first screams louder.
But louder does not necessarily mean more important.
The technology capable of producing a restaurant menu and the technology capable of accelerating scientific discovery are not completely separate historical phenomena. They are emerging from the same extraordinary expansion of machine capability. We don’t get to neatly select the parts that flatter our values and reject the parts that threaten our identities. We cannot simply say: automate protein modeling, accelerate drug discovery, solve fusion, help with climate change, and then please stop immediately before Photoshop.
Technology does not evolve according to the boundaries of our personal comfort.
Which leaves us with an uncomfortable question: how much progress should humanity surrender so that the things I happen to enjoy doing remain economically necessary?
I don’t know exactly what lies beneath the resistance. Probably several things at once. Fear is certainly part of it. If something you’ve spent your adult life learning to do can suddenly be accomplished in seconds by software, fear is not irrational. Economics matters too. It is difficult to talk philosophically about technological abundance when rent is due Friday.
And yes, I think ego is involved.
Creative work is especially complicated because we don’t merely earn money from it; we build identities around it. I am an artist. I am a writer. I am a designer. I am a programmer. Then a machine appears that can participate in the activity through which we distinguished ourselves from other people.
Technology says: this capability may become abundant.
The ego hears: then what makes me special?
Those are not the same statement.
We have spent centuries tying human worth to usefulness. Your labor earns resources. Your expertise earns status. Your scarcity determines economic value. Become unusually good at something difficult and society rewards you precisely because you possess a capability other people lack.
Then intelligence itself begins becoming abundant. Of course that is psychologically destabilizing. But perhaps the problem is not abundance. Perhaps the problem is that we built so much of our identity around scarcity.
There is a distinction I think we desperately need to make as this technology develops. The right of a human being to live a dignified life is not the same thing as the right of their current occupation to remain necessary forever.
We should protect people through technological transitions. We should redesign economic systems when automation destroys the mechanism through which people access food, housing, healthcare, and security. We should fight like hell to ensure that the abundance created by artificial intelligence does not simply become another machine for concentrating wealth. If AI produces unimaginable abundance while a tiny number of companies own everything and everyone else is left begging for access, that is not liberation. It is merely a more technologically impressive form of feudalism.
But preserving unnecessary labor cannot itself be the goal.
Human beings have a moral claim to share in the abundance created when necessity disappears. Human occupations do not have a moral claim to remain necessary.
Those are radically different ideas.
When a machine safely performs some miserable task nobody wanted to do in the first place, we celebrate. When a machine performs something around which we constructed an identity, inefficiency suddenly becomes sacred. Yet the universe does not owe any of us permanent economic relevance, and humanity should not deliberately remain less capable simply so our existing hierarchies of expertise survive intact.
I know the cancer comparison is extreme. That is why I think it matters. A designer complaining about generative AI is not preventing leukemia research. Someone refusing to use ChatGPT is not killing cancer patients. Artificial intelligence is not guaranteed to magically cure cancer, solve hunger, discover limitless energy, or usher us into some frictionless technological paradise. Anyone claiming certainty about those outcomes is selling something.
But extremes are useful because they expose principles.
If the underlying claim is that certain forms of human intellectual or creative activity are so meaningful that machine capability should be intentionally restrained in order to preserve a meaningful human role, then we should be willing to follow that principle all the way through.
Would we accept slower scientific progress?
Would we accept slower drug discovery?
Would we accept slower development of clean energy?
Would we accept hunger lasting longer?
Would we accept disease lasting longer?
Would we accept children suffering longer?
At some point we have to ask whether preserving the meaningfulness of human work matters more than the consequences of the work itself.
For me, the answer is no.
The purpose of medical science is not to provide scientists with meaningful careers. The purpose of medical science is to heal people. The joy scientists receive from discovery is beautiful. The meaning they derive from their work is real. Their curiosity is one of the finest things humanity has ever produced.
But the child is the point.
And perhaps this is the question underneath much of our panic about artificial intelligence. Not simply, What happens when machines can do our jobs? but something considerably more frightening: Who are we when usefulness is no longer the price of admission to existence?
Our civilization has almost no answer for that.
We know how to assign value through scarcity. We know how to determine status through expertise. We know how to distribute survival through employment. We know how to tell someone that they matter because they produce.
We are much worse at saying they matter because they are here.
Artificial intelligence may eventually force us to confront that distinction.
Maybe you will still paint after machines can create astonishing images. People still play chess even though machines surpassed the best human beings decades ago. Maybe you will still write songs. Maybe you will still design menus. Maybe you will garden even after agriculture becomes almost entirely automated. Maybe scientists will still study the universe even if machines can race farther ahead than any individual human mind.
The activity does not become meaningless merely because you are no longer the most efficient entity performing it. Perhaps it becomes something purer. Something you do because you love doing it. Not because scarcity forces another human being to pay you for possessing a capability they lack. That sounds less like the death of creativity to me than its liberation.
I don’t know what artificial intelligence will become. I don’t know whether the enormous promises being made about scientific acceleration will materialize. I don’t know whether AI will produce abundance or whether human beings will find some spectacularly stupid way to use unprecedented abundance to create even greater inequality. History suggests we should at least entertain the possibility.
But I know what I want us optimizing for.
Less suffering. More possibility. More knowledge. More abundance. Longer healthy lives. Children growing up.
And if achieving those things eventually means surrendering some activities through which human beings currently derive money, status, or meaning, then our challenge is not to stop progress in order to preserve necessity. Our challenge is to build a civilization in which losing economic necessity does not mean losing dignity.
I love human creativity. I love art. I love discovery. I love the strange satisfaction of making something that did not exist before. I want people to paint and write and design and experiment and make music and study the universe for as long as human beings exist.
But I do not love any of those things enough to ask someone else’s child to wait longer for medicine so that I can remain necessary. Yesterday morning that was a philosophical position. Then my daughter came home from school. Now it has a face.