Church Anew

View Original

Sermon Writing with AI: When ChatGPT Meets the Lectionary

Photo by Solen Feyissa on Unsplash


I am both a proponent and a skeptic of using artificial intelligence for ministry, and especially for preaching. I believe church leaders must learn to work alongside generative AI to create sermons that are more thoughtful and engaging. I also worry that AI’s opaque algorithms and carefully concealed agendas can lead to sermons that are far removed from core doctrine and long-held theological convictions. Ask ChatGPT to write you a sermon and it’ll access any and all content it can scrape, with little regard for the authority of the source matter or the motivation of the author. That’s why it’s so critical for church leaders and preachers to learn to work alongside AI with equal levels of interest and skepticism, recognizing that there are appropriate and advantageous uses, just as there are applications that are unfaithful or even unethical. Yet one of the ways that preachers might work alongside AI is in reading the lectionary. 

Many preachers begin their sermon preparation by consulting the Revised Common Lectionary for next Sunday’s readings. They’ll read the collection of texts: Old Testament, a Psalm, New Testament, and Gospel. They might then consult a commentary to guide their reflection and interpretation. Some might translate the texts, using a lexicon to expand the possible meanings of the given text.  They might consult a digital resource like a blog or podcast to find the interpretations of trusted thought leaders. Throughout this process, a preacher might struggle to find the overlap or connection between the texts. However well-intentioned the compilers of the lectionary were when they approached their work, it’s typical to find passages in the same week of readings with disparity or even contradiction. Sometimes, these disparities are the result of genre-jumping. Law from the Old Testament meets letters from St. Paul. A Psalm of lament is bolted on to a mundane story from the Gospels. In other cases, these points of divergence arise from the plurality of voices in a collection of texts, whose words were compiled thousands of years apart from one another. As the preacher wades too deeply into any one of these texts, they stray further from any points of thematic overlap between the collected lectionary texts, increasing the likelihood that they’ll cast aside all but one of the readings during their sermon. In other words, the further the preacher wades into any one text, the more likely it becomes that they’ll miss ‌continuity and coherence from the collection. 

This isn’t to say that preachers shouldn’t be careful readers and researchers. Faithfully crafting a sermon tends to require some in-depth engagement with authoritative resources. But faithful sermonizing alongside AI involves identifying the points of overlap between the texts before one wades too deeply into any one reading. With AI, the preacher may want to connect the proverbial dots before researching each text independently, starting with continuity before diving into specificity. Even if the church leader intends to only preach on one text, this process of identifying overlap and coherence expands our imagination of the possible meanings of an individual text. With that in mind, here are 4 ways to read the lectionary alongside ChatGPT. 

Follow along with the free version of ChatGPT. You can access my templated version of this process here.

First, find the texts

ChatGPT can identify the lectionary readings for an upcoming Sunday. To connect the dots across a set of readings, start by prompting AI to find the readings for you. Note that AI won’t always be able to paste the texts within the application due to copyright restrictions. AI companions are also helpful at parsing different translations, identifying similarities and points of convergence. 

Then, describe your context and name your intention

Before you ask Generative AI to create anything, you need to engineer a thoughtful prompt. After locating and reading each text, describe your preaching context and goal. In other words, give AI a persona to adopt as it works with you. Your prompt should include a description of who you are, and what you value — along with who is in your congregation, and what messages might resonate. You might identify that you are an ELCA Pastor preaching on the fourth Sunday in Advent to a small suburban congregation in the northeast United States. Or, you might state that you are a lay preacher, filling in for the senior pastor so they can have a Sunday off before Christmas Eve. You might explain that your congregation expects a connection to a particular liturgical theme or to an ongoing sermon series. You might also describe anything that’s taking place within the worshiping community — changes and transitions, longings and losses, sources of joy and celebration. Don’t hit enter just yet. Because your next step is to:

Ask for points of commonality and overlap

Having described your context and intention, ask the AI to locate possible points of commonality, overlap, or even divergence between the texts. If you have a preaching theme, you might ask AI to identify these relationships with respect to that theme (i.e., what do these 4 texts say about waiting for God to act?). Take these ideas as a starting point for further reading and research. AI is prone to hallucinations and mistruths, and it’s been known to offer an occasional embellishment. 

Identify resources for further research and learning

Finally, use these points of overlap to launch into your sermon preparation. Given the areas of overlap, or even areas of divergence within the collection of readings, ask AI to name some preaching resources. Ask for book ideas, podcast recommendations, scholarly articles, or blog posts related to the texts, and related to the points of overlap within the text. By identifying resources for your further exploration, AI acts as a copilot, rather than a creative replacement. This is always the best way to use AI. ChatGPT pairs with the lectionary when viewed as the jumping-off point for our own prayerful creativity, not when we outsource the creative process. 



See this content in the original post