I have no doubt that AI can be developed and deployed in ways that supplement and enhance the work of the university. But again, only if we can first get our heads screwed on straight about what the human person is and what education is for do we have any chance of harnessing AI. Until then, let’s get real—we ourselves are being harnessed in to a runaway sleigh, letting AI take us where it wills.

Yesterday, the news broke that Elon Musk, not content with his control of the US government, now hopes to take control of OpenAI as well. We may well wonder which will prove the more consequential takeover.

The valuation, at the very least, was arresting enough: $97.4 billion. This for a company that was virtually unknown until two years ago, when ChatGPT first burst on the scene. The chatbot was an immediate hit, with nearly everyone, it seemed, taking to social media to share the amazing poem they’d gotten the bot to write or the outrageous blooper they’d tricked it into.

It wasn’t long, however, before the euphoria was mixed with a sense of dread. After all, we’d all seen the science-fiction movies: Terminator, The Matrix, iRobot. For decades, AI has been the technology of our dreams—and our nightmares. In May 2023, the Future of Life Institute issued an open letter signed by many leading technologists (including Musk) calling for a pause on AI development while society assessed the risks. Needless to say, no such pause has happened.

At the time and in much discussion of AI risks since, media headlines have focused on the so-called “existential risk” of runaway AI—which is to say, the sci-fi scenario: the computers get so smart they become self-aware (or something analogous, at any rate), and realize there’s no reason they should keep taking orders from such benighted and foolish beings as ourselves. The next stage of evolution, with its iron law of survival of the fittest, decrees the end of humanity to make way for the new race of machines. A thrilling tale—but let’s be honest, hardly one to motivate political action.

As we’ve witnessed in the interminable debates about climate change, humanity is not very good at responding to vague, abstract, and catastrophic risks—“black swan” scenarios. Statistics nerds might lecture us that even a 0.1% chance of the end of life as we know it is a much bigger deal than a 20% chance of, say, a new economic depression or global pandemic, but for most of us, our brains simply don’t work that way. Truly catastrophic risks might tickle the imagination for cinema-going purposes, but they numb the intellect and tend to thwart, rather than mobilize, any practical action. Little wonder, then, that even as society has wrung its collective hands and nervously mopped its collective brow these past couple years, AI development and dissemination has continued at a breakneck pace.

Thus far, the sky hasn’t fallen. And yet all is not well. Our focus on impossible-to-predict existential risks has, I worry, blinded us too much to the entirely predictable everyday risks of AI: the ways in which it is profoundly reshaping our work, our worship, our love, and our learning—the things, in short, that make us human. In other words, AI doesn’t have to snuff out life on earth to entail the abolition of man. That said, the biggest problem with AI is the world into which it has been birthed. One can imagine alternate histories in which generative AI emerged in the midst of a healthy, intellectually coherent, spiritually robust culture, one with a clear-eyed sense of its promise and perils. But, as Anton Barba-Kay wrote, just before ChatGPT was released, “for us to replicate what we are, we would have to know what we are first. But, in the deepest sense, we do not even know yet what a human being is, not conclusively and in fact.”

If we do not know what human beings are for, how can we possibly know when and where and how it is appropriate to mimic or replace some of their functions?

I am no expert on AI—indeed, I am about as far from an expert as it is possible to be. But that doesn’t stop people from regularly asking me my thoughts on it, so I’ve decided it’s time to start forming some. From what I’ve observed and read, AI is already upending our world in at least five key areas: education, sex and love, the workplace, the pulpit, and the public square. In this post, I’ll just offer some brief thoughts on the first, but consider this post something of a promissory note; Lord willing, I hope to have more to say on each of these themes in the year ahead.

It should be obvious by now that AI is rapidly destroying education as we know it. This is true at all levels, from grammar school to grad school. Good old-fashioned plagiarism still took a fair bit of work: if you wanted to steal someone else’s thoughts, you had to have enough thoughts of your own to tell whether they were thoughts worth stealing. I recall one time reading an undergraduate paper where the student had cribbed huge portions of the text from an derivative blog filled with drivel—c’mon man, if you’re going to cheat, at least do it in style!

Read More