What Efficiency is For
AI raises the stakes on the difference between creativity and efficiency.
Subscribers who have followed my writing know that I am a fan of efficiency. More specifically, I am always looking for ways to reduce the distractions that take away from the things that matter most to us professionally and personally.
Thus, I don’t have social media apps on my phone. I get almost no notifications on my phone or laptop, except for texts from family and friends. I despise ‘reply all’ emails and obsessively encourage senders to use BCC to keep them from happening.
When it comes to creative work, I relish efficiency only in the sense that I free myself from ‘urgent’ distractions so I can consistently write. But we have to draw a hard line at cutting corners on creative work itself.
Nonfiction writing has value to the extent that it examines new evidence, or examines old evidence in a fresh and imaginative way. There is no particular value in writing that examines old evidence in the same way as earlier writers have done.
Take the life of missionary Adoniram Judson, the subject of my current book project. Obviously there are some elements of his biography that any biographer must repeat, such as his birth and death dates. But there is little value in writing a book on Judson that simply repackages what earlier biographers have written.
Some authors, candidly, have basically produced this kind of derivative book on Judson, presumably because they’re fascinated with Judson and just want to retell his story. Perhaps their publisher figures that any new biography of Judson will sell some copies, whether or not the work adds to our understanding of the man and Christian missions.
The most efficient way to generate such a biography of Judson would be to have AI write it for you. I am not aware of AI software that will write a book based on a single command, but presumably you could patch a book together via multiple requests. As I am writing this, for example, I got Microsoft Copilot to spit out a roughly 12,000-word paper on his early missionary career.
The appeal of this method is obvious: it is phenomenally efficient, and for some students, the clarity and evidence in such a paper exceeds what they could do on their own. (Of course, if you present AI-generated content as your own work, that is plagiarism.) In terms of creativity and historical learning, however, the paper has no value at all. The AI program can add to its source base and correct internal errors, but software programs cannot gain wisdom or encouragement from Judson’s life.
Human readers might gain such benefits from an entirely AI-written book, but AI reports are 100% derivative, and in a nonfiction sense they are mere accumulators of facts (or hallucinations).
My point is that AI is forcing fans of efficiency to reestablish a crystal-clear distinction between efficiency and creativity. We should certainly pursue efficient methods of writing: for example, it would wreck my writing efficiency if I insisted on writing out book manuscripts by hand and not using computers.
But in when writing books, music, or similar endeavors, there is no substitute for a person creatively grappling with the material at hand. Regurgitation by software is not, and won’t ever be, a satisfactory equivalent.
“Reexamining Thomas Jefferson” - my review of three books on the American Founding, at Christianity Today.
“My Top Books of 2025,” at The Gospel Coalition.