3 questions to help us use AI better.

*****

The Preacher declares in Ecclesiastes 1:9 that “there is nothing new under the sun.” He was saying it with a sense of despair, but there’s a reality there that can prove helpful too. Even with something newfangled like AI, the issues aren’t all new. And because we’ve seen some of this before, we can look to history for guidance.

So here are three questions to help us look to the past to see how we can best use today’s AI.

Q. 1: Am I using AI to skip my exercise?

For as long as there have been schools, there’s probably been boys learning how to do division in ways their math teacher wasn’t intending. If little Timmy and a couple of his third grade buddies did their homework together, the three of them would quickly realize that an assignment of 15 questions could be done 3 times as quickly if they did 5 questions each and shared their answers.

That gets the assignment done in record time, but these “get ‘er done” boys misunderstood the point. The teacher’s goal isn’t simply to get 15 right answers from them – she could go to the back of the teacher’s edition if that was all she was after. What she wants is for her students to go through the struggle of working through each of the 15 questions so that their brain muscles will grow. What Timmy has done instead is the equivalent of recruiting his two friends to each attend a third of his basketball practice for him. That’d be a better cheat actually, because it wouldn’t take him long to figure out that his shooting percentage isn’t going to improve so long as he isn’t putting up the shots.

The issue is older even than schools, addressed in various ways throughout Scripture, but maybe most pointedly in Proverbs 10:4: “Lazy hands make for poverty, but diligent hands bring wealth” and Galatians 6:7: “Do not be deceived: God cannot be mocked. A man reaps what he sows.”

Now, in our AI age, there are students using ChatGPT, Claude, and Gemini to skip their school “workouts” in this updated fashion, and still not making the connection as to how that’ll keep them mentally and spiritually flabby, stopping them from growing in knowledge, wisdom, and capacity.

Adults have their own version of this problem, complicated by the reality that sometimes it is just about “getting ‘er done,” while other times it’s about needing to get stretched and refined by the work you put in.

Need a pile of rocks moved? Then having a couple of friends help is a great idea. Got access to some tech that will do the heavy lifting for you? Great, go drive that tractor over here.

In an office setting, AI can help us move all sorts of “rocks,” by doing the big-effort-but-little-thought tasks of quickly compiling your week’s billable hours, assembling your spreadsheets into an array of insightful, colorful pie charts, or gathering and summarizing vast amounts of research material.

But if, like little Timmy, we thought it was always about the results, we’d miss out on when it’s important to put in the sweat. There’s a Stephen King quote about how “I write to find out what I think” and it’s exactly there that AI might be most harming us. We don’t all have to be writers, but if we’re going to be thinkers, then we need to be able to get our thoughts down on paper, or up on a computer, or batted around with our coworkers. That process will allow them to be tested and refined, and sometimes discarded. It’s the sweat invested that helps us really think through things. AI can even be a part of the process, being a part of the batting-around refinement. But we can’t use it to skip over that process.

When we should or shouldn’t use AI will depend on whether we just need to get ‘er done, or whether we should be putting in the exercise to increase our own capacity, knowledge, and wisdom.

Figuring out which we’re doing is complicated by the fact that our jobs are going to regularly involve a mix of both. A pastor might use AI as an upgraded search engine to much more quickly gather up what his most trusted theologians have said about the text he’s planning to preach on. But he can’t pull a Timmy and have AI write his sermon, because his job isn’t simply to read a sermon, but to glorify God in the preparation of it, so that he can tailor it to the flock God has entrusted to him and not to AI.

Q. 2: Am I owning my output?

And that takes us to the issue of responsibility. AI brings up some powerful temptations on this front, but, again, it’s nothing new for Man to try to avoid blame by sidestepping his own responsibility.

“The man said, ‘The woman you put here with me – she gave me some fruit from the tree, and I ate it’” (Gen. 3:12).

With AI’s instantaneous output, what that prompts is the very strong temptation to skip over the “is this actually right?” stage. After all, if you’ve generated a whole report in seconds, doesn’t it seem weird to spend half an hour checking through it? And that’s how you get output like the pictured post from the US Department of Education. It was meant to promote jobs that AI isn’t going to replace any time soon, but it went viral for showing what kind of output AI can provide when the humans in charge aren’t really taking charge. Pipes, pipes, everywhere… and what is that tool she’s using?

In our day-to-day, this temptation will pop up in small ways, like using Gemini’s auto-response to reply to an email you’ve barely read. See how impressed your coworker will be when he pops by your office door for a follow-up and you can’t recall what he’s even talking about because your brain was never engaged. This isn’t a big thing – it might be the difference between having AI complete an email response in 3 seconds that you should have taken 30 seconds to do yourself so you’d know and recall what you wrote.

In our schools, AI can be used to generate math and spelling worksheets, and history and geography pop quizzes in just seconds. For the most part, that’s just AI helping teachers “move rocks,” and what a wonderful resource for them to have. But their students will be wondering why teachers can turn to AI to do the heavy lifting when their pupils aren’t allowed to. To keep students’ hypocrisy-detectors from firing off, teachers will then need to model the highest standards of responsible AI usage. Practically speaking, that means putting in the time proofreading whatever they have AI outputting. How can they convince students they need to put in the sweat if they’re not willing to sweat themselves?

Q. 3: What’s missing from my AI output?

What makes a lot of AI output cheesy? What’s “off” about it?

The other side of this question would be, “What makes something valuable?” There’ll be a subjective nature to any answer to that question because there’s a degree to which beauty is in the eye of the beholder. One person might like Rembrandt, another AI-manufactured alien landscapes. To each their own.

But a friend passed along something he’d learned from a Tim Keller sermon on 1 Cor. 13. Here Paul makes the point that prophecies, tongues, and even charity, without love, are nothing.

While we were excited about all that AI could do just a year ago, now we’ve coined the term “AI slop” and lament the “AI gloss” detectable in its pictures and written work. Maybe what Paul is saying explains this reaction. Someone could use AI to produce a “new” Rembrandt or maybe a “new” piece by Bach, manufactured in seconds by having these artists’ styles convincingly mimicked. The result could be beautiful, and we might not detect even a hint of “AI gloss.” Yet would we treasure these new works like the old?

No.

Why?

Because AI allowed them to be created without sweat or thought. They were made without love.

Love is why the kids’ art on Oma’s fridge is treasured – because of the love going both ways. And love is why the poems many households craft to accompany Christmas presents were better before AI was there to perfect them. Butchered meters and forced rhymes – and the care involved in crafting each line – were what made the merriment.

Could an aspiring poet partner time and intent with AI to lovingly craft a poem for his beloved wife? Certainly. AI usage doesn’t have to be loveless.

But AI without love? That’s nothing at all.

Jon Dykstra realizes he should write his beloved wife more poetry, with or without AI help.