Some of us are reading through C.S. Lewis’ The Four Loves this summer for the Chi Alpha Summer Reading Project. Every other week I’ll post some reflections on the readings.
Today we complete our first reading, the ten pages of chapter 1.
What stood out to me is something that probably seemed like a throwaway observation back in 1960.
I was looking forward to writing some fairly easy panegyrics on the first sort of love and disparagements of the second. And much of what I was going to say still seems to me to be true…. Every time I have tried to think the thing out along those lines I have ended in puzzles and contradictions. The reality is more complicated than I supposed.
Lewis knew what he intended to write, but trying to work it out clearly enough to put it on paper showed him that his thinking was fuzzy. Contradictory, even. Putting feelings, impressions, and assumptions into words is clarifying.
generative AI has entered the chat
ChatGPT and its competitors are tools and they have a place, but please don’t let them undermine your ability to write out a clear argument. Writing what you think is one of the only ways to force yourself to grapple with what you think. Talking it out can also help, but it’s not as brutal as writing. The flow of conversation can allow you to gloss over a weak point in your argument, but having to write out each of your assumptions and inferences on paper doesn’t provide such wiggle room.
I think most of you know that I write my sermons out word-for-word and then try to deliver the sermon without consulting my notes. Why do I write my sermons out if I don’t intend to read the resulting manuscript? It’s for precisely the reasons I mentioned above: to write it out means that any weak spots in my thinking become clear. I still make mistakes in both interpretation and argumentation, but I avoid a lot of obvious mistakes that would otherwise crop up. Delivering the sermon without the notes is about better connecting with the audience. If my thinking on the subject is sufficiently clear, I don’t need the notes except for when I’m quoting a passage from the Bible or some other source.
How does generative AI play into this? I don’t use AI to write my sermons because the goal isn’t a well-written sermon, the goal is a thought-through sermon. And specifically, a thought-through-by-me sermon. A well-written sermon is mostly the byproduct of preparing a well-thought-through sermon. And so if I were to use a tool like chatGPT to write a sermon for me, I would be an actor, not a preacher. Actors need scripts. Preachers need convictions. I need to know (and I need you to know) that I believe what I preach, and I can only know I believe it fully if I write it myself.
Even if I became confident that a ChatGPT sermon would be better than mine and you would enjoy it more, that wouldn’t sway me. Preaching that way would enfeeble me, perhaps even corrupt me. To be a preacher means many things, but among them is the claim that I really mean it. Not just that I mean the things I say in that specific sermon. I have to mean the whole Christianity thing. To be a preacher is to claim that I’m doing my best to follow Jesus. Even if I never preached a sermon against slander, if I had a habit of posting slanderous things on social media you would nonetheless judge me a hypocrite and someone who should be kept away from the pulpit. To stand in the pulpit is to stand before God and man and say, “I really mean it and I’m trying.” Part of that “really meaning it” is manifest in the way I prepare sermons.
This isn’t a new thing. Even before tools like chatGPT came along every preacher had the option of plagiarizing other preachers’ sermons. It has always been looked down upon, partly for its dishonesty (one of the implicit claim of a sermon is “this is what I came up with”) and partly because it meant the preacher wasn’t growing — the act of crafting a sermon makes you a better Christian (or forces you to embrace hypocrisy) and a clearer thinker.
This is not an anti-AI rant. I will sometimes use generative AI after I’ve written my sermon. I will give it prompts like “Here is the manuscript of a sermon I intend to preach to a group of Stanford students. What’s the biggest blind spot in this sermon?” or “What’s the most devastating critique you can make of it?” or “Is there anyone this might needlessly offend?” And then I’ll take that feedback and use it to refine the sermon. Using AI like this is fine because it forces me to strengthen my thinking and wrestle with my convictions. At times the AI has suggested that I should take out a potentially offensive claim or tone down some rhetoric and I’ve thought, “Nah — this is what people need to hear and this is how they need to hear it.” Other times I consider the feedback and say, “Huh — I hadn’t thought about it that way. Yeah, let me reword that so that I’m making the point I intend to make and not being distractingly offensive.”
Obviously, none of you are preachers (at least, none of you has that as a key part of your job). But there is probably some area of your life where you need to be able to think clearly and to know that you have thought clearly. Don’t allow the wonderful tool of generative AI to keep you from developing that skill. If you’d like to mull that over, I recommend the wonderful and very short story The Whispering Earring.
Lewis, of course, had no idea that such a thing as generative AI would ever be invented. He just mentioned that his thinking about love was unclear until he tried to write about it. One of the beauties of reading a well-thought-through book is that it continues to have relevance decades after it was written and that its insights are relevant to new domains that did not exist when its arguments were crafted.
If you’re not reading The Four Loves with us, I highly recommend it. You can download a free copy at archive.org.