Don't Let AI Make You a Lying Thief
Enough is enough
When I was in preaching school, arguably the biggest rule in the student handbook was DO NOT PLAGIARIZE.
We were there to learn to study the Bible and share it with others, and so to download someone else’s sermons and preach them as our own, or to copy and paste a research paper and avoid doing the study ourselves was an egregious sin.
If that’s what we were going to do, there was no point in even attending.
On a recent episode of Think Deeper, I pointed out that if downloading another preacher’s sermons and preaching them without attribution is a fireable offense (as it should be), then AI-generated sermons should receive the same punishment.
Ai-generated articles should receive the same level of skepticism. Churches would be wise to start putting anti-AI clauses in preacher contracts.
Here’s what you have to understand about the “articles” and “sermons” ChatGPT and Grok write:
They don’t write at all. They’re slapping together a conglomeration of the best results from an elaborate Google search.
They’re taking things real writers have already written, combining them without attribution, and handing it to you in a sleek package.
That’s right: AI-generated writing is just plagiarism with extra steps. Plagiarism, of course, is both lying and stealing, as you’re taking someone else’s work from them, and then lying by putting your name on it.
To give you an example - I was researching different views in my study on Revelation and came across a site that made a claim that I found dubious. I researched it some more and decided the author had his facts wrong. When I was done, just for the fun of it, I punched the question in to GPT. It copied that dubious author’s claim almost word for word.
How do I know it’s a problem?
AI detection tools have proven pretty useless. I’ve tested them by writing an article an immediately dropping it in to multiple detectors, only to be told that a large percentage of it was done by AI. That’s no use.
But that doesn’t mean people can’t tell. Spend any time using an AI LLM and you start to see how it talks. Lots of em dashes—you know, these things—is a tell. The outlining can be pretty obvious. Short, repetitive, punchy sentences start to pile up.
More than anything, AI allows an enormous increase in output. It can take hours to think up an article idea, consider how to structure it, write 1,000 words, revise it, and publish it. It takes 5 seconds to enter a prompt specifically requesting a 1,000 word article on a given topic. If you put some effort into shaping the article with the LLM’s help, it might take 5 minutes.
For sermons, it’s a lot harder to tell. There are no verbal em dashes. One preacher recently argued that AI for sermon prep is great, because it frees up time to do all kinds of other ministry work.
But in the long term, people are going to notice. You can’t hide a lack of study, a lack of wrestling with the text.
And, again, AI isn’t actually studying the Bible for you. It’s reproducing someone else’s Bible study. Taking from better men who put the work in and deliberated at length over the text and standing up like you did too is unconscionable. As Forest Antemesaris recently wrote, it’s trading our eventual birthright for a bowl of stew today.
I’m not entirely anti-AI.
If you think of GPT as a fancy Google search, it can save time on some things. But when you start getting it to do your thinking, you’re playing a dangerous game. And more than anything, getting it to steal someone else’s work for you is something we all need to pledge to make entirely off-limits.
This is all still fairly new, so it’s entirely possible a lot of folks don’t know how it works. But that can only be an excuse for so long. Don’t let AI make you a lying thief.
Notes
Grab a copy of You Are Saved today!
