Artificial Intelligence Sermon Generator

[Rev. Edward Engelbrecht]

Ah, gentlemen, if you are using the internet for sermon research such as reading the news, you are already using artificial intelligence to prepare your sermons. It's here now, part of every search engine. It's in your hands. You've already embraced the technology.

This is the irony of technology, I think. We fear it while we use it and depend on it. That personal,  appropriate,  local application you are making is supported by algorithms.

[D. Engebretson]

One downside, I fear, is where pastors would use the AI created 'product' as is just like they did years ago with Concordia Pulpit without any consideration for local context or the need to apply the sermon to a specific audience.  I trust the group would not be large, but it would be there, and congregations would be the poorer for it.

[Charles Austin]

Pastor Engebretson:
One downside, I fear, is where pastors would use the AI created 'product' as is just like they did years ago with Concordia Pulpit without any consideration for local context or the need to apply the sermon to a specific audience.
Me:
Yes. Lazy pastors would be tempted to do that.
There is something different about using the current type of AI from doing your own research, whether that be in commentaries, online, or at the library.

[Matt Hummel]

I think it was on either on Jonah Goldberg’s or Bishop Baron’s podcast where an expert on AI was that executive function decisions are being made for you at levels you do not realize. It is a black box into which you have no access and it provides you the answers it deems appropriate for you.

If you have never watched the series Person of Interest, then do so. AIs are not our friend. And they are a dangerous tool.

Interesting to note that the world of Dune began with humanity’s uprising against computers (the Butlerian Jihad) because AIs were declaring who could give birth and who must abort.

[Brian Stoffregen]

One downside, I fear, is where pastors would use the AI created 'product' as is just like they did years ago with Concordia Pulpit without any consideration for local context or the need to apply the sermon to a specific audience.  I trust the group would not be large, but it would be there, and congregations would be the poorer for it.

A speaker I heard stated that CPH stopped publishing Concordia Pulpit because too many pastors were preaching those sermons from the pulpit as if they were their own. Similarly, an assistant pastor in an LCA congregation told me that he followed along in, I think Augsburg Sermons, word for word the sermon of the senior pastor.

Related to this, how much time each week should pastors spend researching and preparing their sermons? A general rule of thumb I heard was an hour of preparation for each minute of sermon. This means 15-20 hours per week - roughly half a work week on sermon preparation. Granted, visiting people can also be seen as sermon preparation as we learn about the needs and concerns of the people in the pews.

[Charles Austin]

On the other hand, recently, I heard a sermon, and I’m sure I would’ve preferred something that artificial intelligence wrote to the one I actually heard.

[Rev. Edward Engelbrecht]

On the other hand, recently, I heard a sermon, and I’m sure I would’ve preferred something that artificial intelligence wrote to the one I actually heard.

Made me laugh outloud!

I learned the one hour preparation to one minute preaching ratio,  too. I think this is taught by homoletics professors who want to hear a decent sermon from students.

My preparation is not nearly so much per week but my preparation is long. I annually plan all my sermons for the entire year, choosing the text, theme, focus verse, doctrinal summary, and title in the summer for the coming church year. Alongside this, I choose the sermon hymn and bulletin cover. So I know all that before the week's preparation begins, which includes reviewing my prep, forcing an outline or two, and then taking notes as I go about visitation, etc. Reading Study Bible notes, news, and my people usually populates the notes, which I write into a sermon at week's end and too often at 5 AM on Sunday morning.

[D. Engebretson]

I, too, heard the rule of "one hour for one minute".  Must have been the thing in our time.  I think it makes a point that adequate preparation is needed if the sermon is not going to sound 'off the cuff'.  That said, depending on the week and the needs/size of the parish, finding those 15-20 hours might be a challenge, to say the least. 

But it is true that sermon prep also takes place outside the study, as is noted.  The sermon must be contextualized for a given congregation, so the needs and struggles of the people of that place must be in the forefront of the preacher's mind. 

I also consult a variety of sources, sometimes including the sermons of others, such as Selected Sermons of Norman Nagel: From Valparaison to St. Louis (CPH 2004), mainly to see how they treated the text and what themes they may have discovered. Additionally I will consult my own sermons from three years prior when I last preached on the text. I do not always agree with how I did some things, or realize that circumstances have changed, and that launches me off into additional study to see if I possibly misunderstood parts of the text that I could understand better.  For that good commentaries are a must. 

This morning was an interesting day to revisit a prior sermon, as this Sunday, three years ago, was the first Sunday we were in 'lock down' mode in Wisconsin when our governor ordered public buildings to be at a 10 or less capacity. That Sunday was the first time the in-person 'congregation' consisted of only three people; a reality we lived with for two and a half months as I preached to a cell phone week after week.  Historically I had approached the text of the healing of the blind beggar from John 9 in the context of the beginning of a pandemic with all of the fears and questions attendant to that event.  Seeing a merciful God in the midst of suffering still seemed timely, and I revised that sermon with the added perspective of how God brought us through that difficult time. 

I don't think AI could have done all that.   

[Brian Stoffregen]

We should also remember that Ai was the name of a town destroyed by the Israelites (Joshua 7:1-8:29)

[Rev. Edward Engelbrecht]

I, too, heard the rule of "one hour for one minute".  Must have been the thing in our time.  I think it makes a point that adequate preparation is needed if the sermon is not going to sound 'off the cuff'.  That said, depending on the week and the needs/size of the parish, finding those 15-20 hours might be a challenge, to say the least. 

But it is true that sermon prep also takes place outside the study, as is noted.  The sermon must be contextualized for a given congregation, so the needs and struggles of the people of that place must be in the forefront of the preacher's mind. 

I also consult a variety of sources, sometimes including the sermons of others, such as Selected Sermons of Norman Nagel: From Valparaison to St. Louis (CPH 2004), mainly to see how they treated the text and what themes they may have discovered. Additionally I will consult my own sermons from three years prior when I last preached on the text. I do not always agree with how I did some things, or realize that circumstances have changed, and that launches me off into additional study to see if I possibly misunderstood parts of the text that I could understand better.  For that good commentaries are a must. 

This morning was an interesting day to revisit a prior sermon, as this Sunday, three years ago, was the first Sunday we were in 'lock down' mode in Wisconsin when our governor ordered public buildings to be at a 10 or less capacity. That Sunday was the first time the in-person 'congregation' consisted of only three people; a reality we lived with for two and a half months as I preached to a cell phone week after week.  Historically I had approached the text of the healing of the blind beggar from John 9 in the context of the beginning of a pandemic with all of the fears and questions attendant to that event.  Seeing a merciful God in the midst of suffering still seemed timely, and I revised that sermon with the added perspective of how God brought us through that difficult time. 

I don't think AI could have done all that.   

The question might be,  "What could AI do to help the preparation process?"

[D. Engebretson]

I, too, heard the rule of "one hour for one minute".  Must have been the thing in our time.  I think it makes a point that adequate preparation is needed if the sermon is not going to sound 'off the cuff'.  That said, depending on the week and the needs/size of the parish, finding those 15-20 hours might be a challenge, to say the least. 

But it is true that sermon prep also takes place outside the study, as is noted.  The sermon must be contextualized for a given congregation, so the needs and struggles of the people of that place must be in the forefront of the preacher's mind. 

I also consult a variety of sources, sometimes including the sermons of others, such as Selected Sermons of Norman Nagel: From Valparaison to St. Louis (CPH 2004), mainly to see how they treated the text and what themes they may have discovered. Additionally I will consult my own sermons from three years prior when I last preached on the text. I do not always agree with how I did some things, or realize that circumstances have changed, and that launches me off into additional study to see if I possibly misunderstood parts of the text that I could understand better.  For that good commentaries are a must. 

This morning was an interesting day to revisit a prior sermon, as this Sunday, three years ago, was the first Sunday we were in 'lock down' mode in Wisconsin when our governor ordered public buildings to be at a 10 or less capacity. That Sunday was the first time the in-person 'congregation' consisted of only three people; a reality we lived with for two and a half months as I preached to a cell phone week after week.  Historically I had approached the text of the healing of the blind beggar from John 9 in the context of the beginning of a pandemic with all of the fears and questions attendant to that event.  Seeing a merciful God in the midst of suffering still seemed timely, and I revised that sermon with the added perspective of how God brought us through that difficult time. 

I don't think AI could have done all that.   

The question might be,  "What could AI do to help the preparation process?"

I'm not sure, since I have no real experience with it. I really want to know more about how AI generates its results; that is, the source material it uses.  I'm under the impression that it gleans from the internet from a variety of sources.  Does it document those sources? I think before I would use it, I would like to understand better how it works.  But it could be a possibility.

[peter_speckhard]

No matter where the words came from, it matters for the purpose of proclamation that the preacher take accountability for saying them. He is declaring something in the Name of God and from his office as a servant of the Word. AI can possibly come up with a good sermon text, but cannot preach it or listen to it. 

[D. Engebretson]

That is why I would want to know the source material AI used.  I agree, the preacher is accountable alone for what he preaches.  We must make choices in information we choose to include or exclude from our sermons.  AI is not bound to this responsibility in what it chooses to include in a so-called sermon.  At the very least I would like to see it document its sources.

[Rev. Edward Engelbrecht]

The internet search engines you use are all examples of AI. They assess your search request and then prioritize the results based on what AI thinks you want. An AI program designed to support sermon preparation could be programed to display the sources of information. The newer feature of AI is the ability compose, as I understand it. If it composed a paragraph,  I suppose it is programmable to tell you the basis of its composition.

[MaddogLutheran]

I think it was on either on Jonah Goldberg’s or Bishop Baron’s podcast where an expert on AI was that executive function decisions are being made for you at levels you do not realize. It is a black box into which you have no access and it provides you the answers it deems appropriate for you.

If you have never watched the series Person of Interest, then do so. AIs are not our friend. And they are a dangerous tool.

Interesting to note that the world of Dune began with humanity’s uprising against computers (the Butlerian Jihad) because AIs were declaring who could give birth and who must abort.

Person of Interest is one of the greatest shows in the history of television.  Both for its compelling storytelling with character arcs beautifully drawn and acted, and the prescient commentary about the benefits and pitfalls of technology interacting with an unchanging human nature.  The longing for connection, isolation, and even eternity/legacy, something outlasting you.

Particularly season 4 episode 5 ("Prophets") which showed (via flashbacks) a few of the many failed iterations Finch went through when attempting to create The Machine (the benevolent AI identifying each week's person of interest).  One of those versions immediately tried to kill him.

The showrunners did a great job anticipating the difficulty in training an AI, as recent examples of a Microsoft prototype going full sociopathic validate.  FYI the malevolent AI on the show was named Samaritan.

P.S. The reason I call it one of the greatest TV shows ever is because it stuck the landing.  It had a destination, it got there, and it made sense where both the story and the characters ended up.  Unlike a lot of shows (glaring disapprovingly at Lost for example).