From our Resource Library

The Digital Pen: AI in Christian Publishing

Learn how to use—or not use—AI in your writing and publishing, in ways that honor Christ.

We’re here to talk about something very close to all of our hearts: the changing landscape of Christian publishing, writing, and editing.

We find ourselves at a very fascinating moment in the entire existence of the written word. We aren’t here today to look at AI as a magic wand or a monster. We’re going to look at it through a much older and much more necessary and biblical lens—and that’s the lens of stewardship.

Language is a profound and uniquely gift given by the Lord. So how do we engage with technology that shifts by the day while protecting the soul, the truth, and the integrity of our words? And that’s what we’ll be looking at.

The paradox of publishing today

So let’s look at first of all the core, the central paradox of today’s publishing.

If I asked you, for example, how many of us use AI in writing, in editing, and publishing processes? I’m sure a lot of us would raise our hands in agreement that we are already using AI. In fact, most of the images that you see, you will be seeing in this presentation today, I have created using AI.

Let’s start by looking at what’s actually happening across the publishing industry as a whole.

The latest book industry study group data shows us something incredible. Nearly half of us are already using these tools in some capacity, and about the same percentage of our organizations, publishing houses, have integrated AI into them.

But if you read between the lines of that data, there’s a collective uneasiness. The numbers tell us that we are adopting these tools far faster than we are becoming comfortable with them. And we feel a natural hesitation, I’m sure. And honestly, I feel that’s a good thing.

It means we care deeply about what we do. It means we know that a machine or AI can mimic a sentence, but it cannot share what comes from the creativeness that God has blessed us with.

Finding a God-honoring way forward amidst the tension

Our goal is to look at the tension together, and to find a grounded, principled, God-honoring way forward. What does this mean for us?

Let’s think of it in two audiences, maybe one, the megaphone, and the other one is the gatekeeper. This tension that I was talking about splits our community into two different roles. On one side, we have many writers and publishers who are often using AI as a megaphone. What are they trying to do? They’re trying to handle the sheer volume of marketing copy of titles and metadata.

On the other side, our libraries, our editors, copy editors are stepping up as gatekeepers. I can say that as an editor, the editing community is deeply concerned about an overwhelming flood of low quality or even shallow machine content that’s entering into the spaces that’s meant for learning and discovery.

When we look at this as stewards, we realize that our job isn’t just to produce more and to take care of what has been produced. Our responsibility is to take care of the entire system, ensuring that whatever we send out into the world is worthy of a reader’s time, attention, and trust.

More than all of these things, does our using of AI bring glory to God? That’s an important question.

AI is a compass and not the captain. So how do we navigate this new terrain without losing our way?

Let’s use a simple rule of thumb: technology can be an amazing compass, but it is a terrible captain. A compass can show you directions, it can lay out structures and linguistic patterns, but it has no idea why the journey matters. It completely lacks discernment, both the spiritual and emotional wisdom to know what is true, what is good, and perhaps what will heal or comfort a reader’s heart.

The moment we hand over the captain’s wheel over to an algorithm, our work loses the orientation, the true north that we need to have.

To keep our focus steady, let’s look at a practical traffic light boundary system that we can use in our daily work. Think of this as a gentle way to protect your creative integrity.

The green light tasks are all about the heavy lifting of organization. For example, if you’re staring at a blank page, it’s perfectly fine to let a model help you sort a rough outline, organize complex data, or even to polish up or clean up basic grammar.

The yellow light is where we pause and we tread a little more carefully. These models are designed to predict the next likely word, but not to know the truth, which means they fabricate facts easily. Even more subtle, they tend to smooth out language until it sounds very average and sterile. It strips away your unique perspective, your personal testimony, and our distinct cultural favor.

That brings us to the red light. Remember, we should never, we must never outsource the actual soul of the message that is being communicated or pretend that a machine’s output is our original unfiltered thought. To do so is cutting corners. It’s a moral compromise. It’s a quiet stepping away from truthfulness. As editors, it’s becoming increasingly difficult to differentiate between human writing and AI writing.

So how do we overcome this? Experience in reading books and articles written by human beings is perhaps the best way to detect AI writing. In fact, plagiarism is now eclipsed by entire books, dissertations, blogs, and yes, even sermons being written from a scratch by AI.

Tools to consider

As believers in the Lord Jesus Christ, we need to seek God’s discernment when we edit, when we write, or publish. Keeping that in mind, there are three AI tools to consider for writing and editing assistance.

ChatGPT serves as an all-round writing and publishing assistant.

It’s good for brainstorming, for outlining, for editing and refining ideas. It can help you develop and improve your work while you are firmly in control or should be firmly in control of the thinking, of the theology, and you should hold the final responsibility as the author.

Gemini is good for everyday productivity. It’s good for research. It’s good for drafting, for summarizing. It’s good because it works well with the Google workspace. It integrates well with Google Docs and Gmail and Drive, sometimes very scaringly, and it makes a good practical assistant. (This does not mean that it should replace human discernment and fact checking. Gemini can easily overstep boundaries and drift into your personal space and make personal observations and even find out details about you.)

Something that many editors use is Grammarly. We use it to correct grammar, spelling, punctuation and improve overall readability. It helps polish the writing without changing the ownership of the ideas. It will not change the cadence and the voice that the reader can expect from the writer.

When and when not to use AI

How do these AI tools help us with the whole gamut of publishing from the time you start writing a book to publishing? AI tools can be used in different stages of book writing, editing, and publishing. AI can help with the research. It can help with the editing. It can help with organization. It can help with design.

But the thinking, the theology, the discernment, and the absolute final responsibility should always remain with the writer. And the editor and the publisher are duty bound to make sure that this happens. This is where faithful stewardship and integrity meet.

There are AI tools for writing and editing and publishing assistance right from the time you have an idea, to research, to writing, to editing, to fact-checking, and of course, design, and even marketing.

Reader-centered prompts

When do we choose to use tools? Our approach needs to be incredibly intentional.

We should practice reader-centered prompts that focus on preserving the unique voice, the cadence, and tone of the author, and how this can be further enhanced but never replaced. We don’t ask AI to create out of nothing. That’s very important. We are, in a way, teaching AI to operate only within and respect clear constraints, protect the author’s authentic voice, and focus entirely on serving the person on the other side of the page by improving the quality of the text, that is the grammar, the syntax, the clarity of thought, and definitely the preciseness. If we type lazy detached prompts, we are going to get lazy detached responses.

Recommended workflow

How do we connect with readers in the age of AI? Here’s a workflow that I would recommend when building out a project or helping an author refine their thoughts.

First, we use a ministry-focused reader profile. We ask AI to look at our core ideas and help us map out who our reader truly is. Ask these questions: What keeps them up at night? What are they hoping for? How can we serve them?

Second, we move to the message clarity blueprint. This keeps us humble. It keeps us focused, distilling the entire project down to into a single, clear, beautiful sentence and a hype-free promise to the reader.

Finally, we need to run the content stewardship audit. We look back over the draft as writers, editors, and publishers to ensure three things. Ask three questions:

  • Is our core message completely true to the word of God?
  • Is the text speaking directly to the spiritual realities of the readers?
  • Are we leaving them with a clear, life-giving next step that honors their sanctification and points them to Christ’s glory?

Here’s perhaps the ultimate standard that we need to remember: Let’s anchor everything that we’ve spoken about today in a timeless standard of craft.

Psalm 19:14 offers a prayer that every writer, editor, and publisher can carry in our hearts: Let the words of my mouth and the meditation of my heart be acceptable in your sight, O Lord, my rock and my redeemer.

The steadiness of our calling

Our systems will change. Algorithms will keep updating week after week, and industry lines will keep climbing charts. But our true calling remains absolutely steady. We are called to be faithful stewards of the words we are giving. We need to ensure that our words are crafted with a clean conscience, with empathy, and an unwavering commitment to truth—the truth that is Christ our Lord.

“Is my use of AI glorifying our Lord Jesus Christ?” I think that’s the central question we need to ask every time we edit, write or publish using AI. If our answer to that question is a hesitation, we don’t go there. But if we’re clear in our conscience about answering that particular question, I think we’ll be safe in how we use AI.

How about other AI tools like Claude, Notion AI, or Perplexity, or other ones? Do you have any input about how those can be incorporated into an editorial process?

Honestly, I have not used Claude or Perplexity or other AI models. I have used mainly Gemini, Grammarly and ChatGPT. I would not like to comment on things that I haven’t used.

Have you ever used AI for translation? And if so, what are some things to be aware of in that area?

I have used AI to translate into some Indian languages, especially a language I’m learning. What I have found is if AI doesn’t have enough data on that language, it mistranslates, so we do need to be careful.

If there is a large enough data set for AI to use—if it’s a popular language or a language that’s spoken by more people, I think it has access to more tools. But I think human checking is very, very important.

Is Grammarly good in all languages or what languages does Grammarly focus on?

I have focused on using Grammarly only in English. In English, I think Grammarly is a very good tool to use, but I have not used Grammarly in other languages.

In India, the Bible Society of India has experimented with AI in Bible translation, and they have found an accuracy rate of 90%. But at the same time, the human element cannot be removed from that process at all. In fact, it should not be. 10% is a significant lack of understanding as well.

Could you comment on the use of AI-generated summary of a concept in one’s manuscript, for example, a reframing of a Puritan’s quote in modern English? Is it okay to use it when citing it in your footnote or endnote?

It should be okay as long as you make sure that it’s acknowledged that it is AI translated or AI summarized by AI. As long as you acknowledge it, I think people should be aware that you have used AI, and let your reader decide whether they’re okay with that or not.

Have you used AI to produce questions or study guides to your original source material?

I have done that for Bible studies, yes.

How do we give proper credit to the use of AI in our writing?

If you noticed my presentation, wherever I have used AI to create an image, I have tagged it and said generated by AI or used AI has been used. We should do it. In your introduction, if you mention that you have used AI, that will communicate your integrity to your reader.

What is the long-term trade-off with increasing our use of AI? Why should we avoid using AI to do thinking for us? (What are the risks versus rewards?)

Two things to think of: Creativity and original thinking has to rest with the human. We can never give that p. The mundane tasks—outlining, summarizing, even abridgement—are things AI can help with. AI in our generation has come to stay.

How do we help our students to be faithful in research and use their brains? That’s the bane of education now.

One of the things I do is edit PhD dissertations. Over the past two years, at least four or five PhD candidates have not graduated because of dishonest use of AI. That is sad. Students must be trained in the proper use of AI, in the ethical use of AI, I must say.

Did I hear right when you mentioned that AI can fabricate quotes, authors, and pages?

Yes. In fact, I did this as an exercise with ChatGPT a few months before I did an editing seminar with MAI in India. Lo and behold, it came up with a completely fictitious reference, author, publisher, and page number. I went back and checked it. When I questioned ChatGPT about it, it said, “I’m sorry, I misunderstood.”

Is your work considered credible if AI is used?

That’s a very interesting and a difficult question because whether we like it or not, AI is here to stay. It’s here to stay even in Christian publishing, writing, and editing.

We must learn to acknowledge AI where it has been used and make sure that it has been used properly. It has not been used to generate the content. It has only been there to support it. Please acknowledge wherever you can. It is a form of compromise and even cheating if we do not acknowledge it.

On strict editing programs such as Grammarly, do you acknowledge its use or do you utilize it much like human editors who are uncredited?

As long as you have not asked AI to alter the content of your writing and you have only used it for grammar and for the mundane tasks, it’s okay to treat it like a human editor. We don’t get credited. In the 40+books I have edited, my name appears only twice.

I had an experience where AI credited me with writing a book, giving it a title, but I never wrote that book. I had written a similar book and it took that information and invented a new book. What do we do about this?

I think you need to make sure that once you have used AI to also ask it to clear your data. Sometimes it can build an entire portfolio of you and very soon access other aspects of you. AI can get to know your wife’s name, your children’s name.

Make sure that you keep deleting your history with ChatGPT and with Gemini. I have a professional account with ChatGPT and it will allow me to erase my history.

How do I as a reader detect incorrect use of AI?

That is a tough one. To spot AI-generated text, perhaps look for a repetitive reliance on certain words. Some things that I have noticed is the word moreover, furthermore. It uses words we don’t normally use. Paragraphs that are highly fluent right in the bang middle of a body of text which does not match that level of language. Look for a lack of depth.

Look for a lack of the human voice. Perhaps the most simplest version would be go to the document’s version history. You can see where massive copy and paste things have happened. Go to the source of the citations. Explore, go and find out if that book is actually there.

How can you clear Gemini if it’s connected with your Google account? Is there a way to do that?

Honestly, I have not explored this. This is something that perhaps even I will explore.

Any perspective on Copilot?

I have not used it, so I cannot comment on what I have not used. I’m sorry.

Do you use different AI platforms to check or critique others’ work?

I’ve not done that, so perhaps it would be an interesting idea.

Would you recommend I use Gemini as an editor and it won’t alter my voice as you mentioned earlier?

I think your prompt has to be as clear and detailed as possible. If you make sure that your prompt tells the AI to make sure that you don’t alter the voice, don’t change the language, don’t change the perspective, it will remain within the constraints. Otherwise, it will go all over the place. Make sure that you spend a lot of time on your prompt to make sure that it doesn’t change your work.

Do you have any sources that you go to for examples of good prompts?

So far, I’ve worked on my prompts alone. I take time to frame my prompts in Microsoft Word before I input it into either Gemini or ChatGPT. Sometimes my prompts are a page long.

What are some of the different qualities of those prompts? What are some of the qualities of a page long prompt?

When I need a work to be summarized so that I get a grasp of what the writer is trying to say, I read the text first. I read the text first, but I keep the summary so that it helps me chapter by chapter, then I go back to it. I limit my prompt to take this text and I want you to do A, B, C, D, E, and I don’t want you to do this, this, this, and this. And with that, if you have a clear boundary, AI can be helpful.

You can have projects with specific instructions that apply to every project, so you don’t need to repeat so much of your prompt each time, because it remembers your approved parameters.

Any further thoughts on how to avoid misusing AI?

Do a self-audit of how much AI you are using. If it’s the first thing you think of when you’re starting a task, that means you’ve already crossed a line. Make sure that you do your work first, whether it’s as a writer or as an editor—if you have not read the work that you are going to edit, if you have not written the work that you are asking AI to edit for you or to refine for you, then I think your reliance on AI is at an unhealthy level. Try working without AI for a while, and don’t use it as the first instance thing.

It’s similar to how we learn languages. If we have learned Greek, the temptation is to reach out to the dictionary first, rather than learning and keeping it in memory. I think these are some things that we need to definitely keep in mind.

I was at a publishing AI writing training back about a month ago, and the presenter asked the room, how many phone numbers can you remember? Then asked, how many of you can remember more than two phone numbers? And there were not very many hands. I remember when I was growing up, I knew lots of phone numbers, but now I don’t know very many, because they’re all on my phone.

What can you say about the ethical implications of the impact of AI use on the environment?

One way, perhaps, to deal with this is to use AI only when it’s absolutely necessary. Treat AI as a very expensive assistant, rather than bombard AI with frivolous questions. If more of us rely on upgrading our own skills before seeking the path of least assistance—that is, AI—I’m sure we can go a long way in addressing the adverse effects of AI on the environment. At the same time, AI is here to stay, so we can’t avoid it, but we can definitely steward it.

Défiler vers le haut