AI cannot read and it cannot write because it is not conscious, and thus it has neither will nor discernment. It believes nothing, feels nothing, and knows nothing. It is we human beings, made in God’s image (Genesis 1:26–27), who read and write because we are conscious agents who desire knowledge and its benefits—at least when we are at our best. AI scans data and produces text. It is a machine. Some of it is impressive and some of it is even true. But we will not let that impress us as writers or replace us as writers.
We refuse it. The machine will not manacle our minds. We will never affix our names to text that we did not write. We will not outsource our personalities to impersonal programs, no matter how efficient or convincing or popular or profitable. We will not bear false witness (Exodus 20:16). We will not trust AI “research” if we cannot verify the source. We will disable or ignore AI suggestions on improving style. We will not let AI “edit” our work, but will rely on editors. We will interact with them on our terms. We will not use AI to “clarify” or to “sound more academic” or to “sound less academic” or to “write at a graduate school level” or for anything. We will not use AI to merge or edit documents. We will do research for ourselves. We will not use AI to save time if in saving time we lose our literary souls to AI.
We will teach others these commitments by precept and by example. As teachers, we will ban AI in the writing of our student’s papers. We will not use AI to grade students papers, since it knows nothing and can write nothing. We will evaluate our students work for ourselves. We yearn to teach students to develop their human potential in humane and personal ways, not to become technocratic text generators. We leave that to those who sell their souls to the Machine, God help them. Read more»Douglas Groothuis | “We Refuse it: A Manifesto on Generative AI Writing” | July 15, 2026
RESOURCES
- Subscribe to The Heidelblog!
- Download the HeidelApp on Apple App Store or Google Play
- Browse the Heidelshop!
- The Heidelblog Resource Page
- Heidelmedia Resources
- The Ecumenical Creeds
- The Reformed Confessions
- The Heidelberg Catechism
- The Heidelberg Catechism: A Historical, Theological, & Pastoral Commentary (Lexham Press, 2025)
- Recovering the Reformed Confession (Phillipsburg: P&R Publishing, 2008)
- Why I Am A Christian
- What Must A Christian Believe?
- Heidelblog Contributors
- Support Heidelmedia: use the donate button or send a check to
Heidelberg Reformation Association
1637 E. Valley Parkway #391
Escondido CA 92027
USA
The HRA is a 501(c)(3) non-profit organization
Heidelblog
The Heidelblog has been in publication since 2007. It is devoted to recovering the Reformed confession and to helping others discover Reformed theology, piety, and practice.
More by Heidelblog ›
I think this guy goes too far. Naturally, having AI author something which one represents as his own is dishonest. but to write something and have AI proof-read it and edit it, polish it up, I don’t think there’s any moral component to that like there is in my latter example. Therefore it seeks like you wouid just want the job done as well as it could possibly be done. if the human editor can do a better job and help you come up with a more polished final product, great. if the AI can, that’s great too. same with AI style prompts. if AI has a great big library full of Strunk and White-type style conventions and recommends them to you from time to time, what’s the harm? using that tool does not compromise one’s ability to honestly represent the work as his own any more than work that’s touched over by a human editor. Ses like if a person is completely honest about turning his work in as his own work, he is the creator of that work, then his consideration ought to be “what source of additional editing will help me get to the highest quality finished project.
This guy probably does his math problems out in long-hand and refuses to use a calculator.
Paul,
Groothuis is an author with a distinguished career. I agree with you, I don’t see the harm in the modest use of AI prompts. I would not turn over my work to AI to revise, editor, or proof without human oversight and correction. Spellcheck is essentially AI and I’ve been using it for a long time but I take his main point, that there is a human component to art and AI isn’t human. It harvests, it recycles, it does what it was programmed to do.
The use of AI prompts, however, will prove to be damaging to young people who will likely never learn to do these things themselves. They probably should learn to do math long-hand before they’re allowed take the shortcut. We’re in crisis in higher ed right now and it’s getting worse every year. Students are entering college/university with a grade-school ability to read and write. University profs are having to teach very basic remedial courses at elite schools (e.g., Cal Berkeley) because student are arriving with such poor preparation. Were it up to me, I would institute a one-year prep course for seminaries. It’s hard to overstate how inadequate secondary and post-secondary ed is in this country.
AI isn’t going to help those students. At best it will mask the deficiencies but that bill will come due.
Dr. Clark, I do get it that for a professionally trained writer like you, AI can be helpful in editing and proofreading. It can catch errors that our human minds not so much overlook, but “mentally correct” in which we read what SHOULD be there, not what actually IS there on the page.
But did you really mean to say that you “would turn over my work to AI to revise, editor, or proof without human oversight and correction”?
Even on your example of spellchecking, I routinely have my spellcheck try to turn relatively rare theological or political terms into something that is more commonly used, but wrong.
On higher-order work, “AI” is better described as “automated ignorance” than as “artificial intelligence.”
I can think of no situation — zero, zip, none, nada — in which I would allow something “proofed” by AI to see the light of day without human review. I have seen far too many cases of AI generating gross errors, particularly on niche topics for which not a lot has been written and posted online, which applies to many theological topics, particularly in the confessionally Reformed rather than broadly evangelical world.
The wife of one of our former state elected officials gave a TED talk recently in which she cited an example of asking AI how many rocks a person should eat daily, and AI came up with the response that everybody should eat at least one rock daily due to mineral content.
Hello? Knock, knock? Is anyone in the machine?
Guess not, based on answers like that.
Darrell,
I meant to type “not.” Typo corrected. I don’t think AI would have helped me there.
Clarification appreciated. That makes a lot more sense now.
One of the problems I have speaking about AI, is that our language makes it difficult to speak about AI accurately without a great deal of circumlocution. Using Alexa as an example, I say “Alexa told me this,” as if there were a person name Alexa who responded to me. But there is no person there. Or I say, “it told me this.” But there is not even an “it” there, in the sense of a single “thing” responding to me.
And if I try to think things out and say what I mean more accurately, I find that I’m tying myself up in linguistic knots, because our language forces me to speak as if I’m assuming something that isn’t really true.
We almost have to use anthropomorphic language, but it’s so easy to forget that’s all it is.
Robert,
We use pronouns for efficiency. I use “she” for Siri because I chose to give it a female (Northern Irish dialect) voice. It’s obviously not female. It’s not human. It’s just a computer that someone (possibly female) programmed. Perhaps I should say “it,” which would ironic since we speak of IT.
100% agree! I am on a quest to find biblical evidence against AI so that I can give good reason as to why it should not be used, and this helps! AI is totally evil.
Oliver,
FWIW, I wouldn’t say that AI is “totally evil” but we should cautious and, in this case, writers should do their own work.
AI covers a lot of ground. When I ask Siri for this or that (the weather, traffic conditions) I’m using AI. That’s not evil. It’s merely harvesting what’s available on the web.
But turning over writing (and perhaps even editing) to AI is to remove the human from the process of creation.