People all over the world are using generative artificial intelligence (GenAI). ChatGPT is now the fifth most popular website, just behind Instagram and ahead of X.1
But should preachers? And if so, how?
To organize my thoughts, I will use a quote I often use when working with students and pastors:
“Technology is neither good nor bad; nor is it neutral.” —Melvin Kranzberg2
AI is not good…
Let’s just say the quiet part out loud: None of us should use GenAI to write sermons for us. As tempting as it may be, entrusting AI to prepare a message is risky for several reasons. Here are three:
- Bias: Researchers are noticing the embedded bias of chatbots. GenAI bias exists for many reasons. One reason is that GenAI learning has mostly been centered on Western views and the English language. A related issue is that this training is based on content created for the internet, which already reflects bias. AI language models have “racist, sexist, ageist, ableist, homophobic, antisemitic, xenophobic, deceitful, derogatory, culturally insensitive, hostile, and other forms of adverse content.”[3] AI programs do not intend to spout out derogatory content. They are not conscious of what they say. But programmers trained them on biased content. Programmers are trying to address the toxic bias that has infiltrated their chatbot training, but the damage has already been done. GenAI perpetuates bias and does so while presenting as a non-biased, efficient machine!
- Hallucinations: GenAI tends to make up information or “hallucinate.” Even the terms of use from ChatGPT tell the user to fact-check the output. Because chatbots mimic our language, we tend to trust that any information they feed us is accurate or factual. However, GenAI language programming is not equipped to understand as we understand—that is, through soft skills of emotional and intercultural competence or interpreting contextual settings. Output is based on what statistically is the most probable sequence of words in response to the words in our prompt. Probability is not necessarily reality.
- Dislocation: GenAI will not produce sermon content through experience, social location, testimony, and personality. Likewise, it cannot understand and adapt content for particular contexts, locations, and situations. This is why, when playing with GenAI for sermon writing, the results tend to be so bland, especially when we haven’t explored the art of crafting prompts that highlight particularity. There is no sense of a particular congregation or situation the sermon seeks to address; a personality is not offering the proclamation. It sounds like an “anybody” message to “generic” people. Boring!
AI is not bad…
If we aren’t using GenAI to write a sermon for us, what else can it do? A lot! Here are a few examples of how chatbots can be prompted to respond to our original sermonic material for other purposes.
- Pre-sermon brainstorming: One way to play with the limits of GenAI noted above is to go ahead and see what the basic, generic sermon on your scripture text would be—and subvert it. Ask a chatbot (like ChatGPT or Perplexity.ai) to write a sermon on this week’s pericope. Chances are, you will see a three-point sermon with a prayer appear on screen in under 30 seconds. Read it and say to yourself: “I can do better than that!” Add the richness of conversation partners, experience, testimony, and contextuality to your sermon. Embrace organic homiletic intelligence!
- Proofreading/listening: Manuscript preachers can play with GenAI to simplify and organize a cluttered first draft. Ask a chatbot to create an outline of your draft, to highlight clunky parts of the sermon, or to summarize the message with a focus statement. If you preach without notes, you can still use a platform like Descript or Otter.ai to create an audio transcript or recording as you talk out parts of the sermon during the week. Play back your recording so you can clean up muddy parts. Ask it to create an outline of your clips so you have an intentional flow on Sunday.
- Post-sermon sharing: GenAI can help us repurpose a sermon for other outlets. One of my favorite ways to play with chatbots is to ask for a series of social media posts based on my sermon manuscript. It will not only capture quotes from the sermon but also create hashtags and emojis for the posts! Platforms like Sermon Shots and Opus can turn a video of your sermon into bite-sized social media morsels to scatter throughout your community’s feeds all week.
AI is not neutral…
One final invitation to you, working preachers: Think about addressing the rise of artificial intelligence in your sermons. Theologically process topics such as: What is sacred about human intelligence or, to be more inclusive, organic intelligence? What is the value of human labor? What is the impact of GenAI data centers on our planet?[4] What is unique about human creativity? Can God work through artificial intelligence? These questions and more should find their way into our proclamation in the second quarter of the 21st century.
Notes
- “Top Websites Ranking,” Similarweb, https://www.similarweb.com/top-websites/, accessed June 4, 2025.
- Melvin Kranzberg, “Technology and History: ‘Kranzberg’s Laws,’” in Technology and Culture 27, no. 3 (July 1986): 545.
- Glenn Kleiman, “Teaching Students to Write with AI: The SPACE Framework,” Medium, January 5, 2023, https://medium.com/the-generator/teaching-students-to-write-with-ai-the-space-framework-f10003ec48bc.
- James O’Donnell and Casey Crownhart, “We Did the Math on AI’s Energy Footprint. Here’s the Story You Haven’t Heard,” MIT Technology Review, May 20, 2025, https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech/.