Some thoughts on the Sermon Assistant from a Catholic perspective

At this time, I use the homily builder solely for lectionary based Bible studies. The Sermon Assistant so far does not impress me for homilies. In short, it goes against my basic understanding of a homily - the breaking open of scripture to feed the assembly. The Sermon Assistant is using artificial human intelligence to assist in a process that should be driven the Holy Spirit and prayer. However, the Sermon Assistant does have significant possibilities as a Bible study assistant - a different focus for marketing.

However, I still find the tools biased in ways that are not helpful to the ACELO churches (high Anglican, Catholic, Eastern Orthodox, high Lutheran, and Oriental Orthodox + some mainline Protestant churches).

the liturgical calendar should be factored into the AI algorithm, as a minimum seasonally. There are a number of books providing good descriptions of the calendar that could be used to inform the AI
quotations could be drawn from collections of quotations and selected by AI. The catalogue of such quotes should allow for user choice in very broad strokes.To be blunt quotations from a Reformed approach are rarely useful to me; Lutheran quotes may be.
at least in my testing, Outlines tend to provide Reformed theology - again the ACELO, pietist, anabaptist perspectives are directly or indirectly of use; Reformed is not.
in illustrations, I personally prefer to use teaching tales from religions around the world or exempla from Medieval Europe, Jewish folklore, or apocryphal stories. ... this use is outside what this tool is trying to do. So I put this feature in the "useful to some, but that some does not include me" class.
in questions, I really need to be able to ask closed questions i.e. questions that have specific answers vs. open questions i.e. questions that encourage discussion. Most of the time the questions I am offered have reasonable quality - occasionally my topic results in rather generic questions that I wouldn't use.
in applications, this section has possibilities, but it assumes a hermeneutical stance re:application that is foreign to me. I look for the literal meaning and one or more of the following (a) something I am to believe (b) something I am to do (c) something I am to hope for (yes - the traditional 4 senses of scripture).

All of which is to say, I think you are heading in a very useful direction - better suited to lessons than to homilies - but you need to get specific feedback from a representative sampling of denominations - Mennonite, Methodist, Presbyterian, evangelical, Stone-Campell, Catholic, Orthodox, Anglican, Lutheran (ELCA/LCMS), Orthodox, Uniting- would cover a big chunk of Christianity. The groups I listed are ones that I know enough about to be able to see how they would perceive the tools' outputs differently than I.

Edit: I should make it clear that there are Reformed authors I respect and that there are other strains of Christianity that are even less useful. But I thought the forumites would recognize the gulf between Catholicism and Calvin better than other gulfs. I mean how many of us truly understand the Waldensian/Catholic gulf?

Find more posts tagged with

Comments

Frank Sauer

While I personally do not fall into the ACELO group, your points highlight a lot of my concern with AI appearing to be the cog in many Features for the next release.

I know that AI is the hot thing, but as illustrated by your experience - AI can definitely come with slanted responses (intentionally or not).

I also stand in the camp of preaching/teaching being birthed out of prayer and directed by the Holy Spirit.

I could see a use for AI in a realm of something like Church History. asking for a view of Church History from each of a Catholic, Reformed and Eastern view would possibly give a nice starting point for study. Hopefully, with accurate presentations of each view, but overall, the process does concern me.

It seems "Logos 11" is being driven towards the entry level user in many ways, with a goal of making it easier - which may be why AI is the dominant base for new features.... But, for new users, especially if they are newer believers - I would be very concerned about AI guiding their study and growth....

David Thomas

a process that should be driven the Holy Spirit and prayer

While I am not in the ACELO camp, I intentionally place prayer in multiple places of my Sermon Prep WORKFLOW because I concur with your understanding of the homily being an exposition driven by exegesis of a text of Scripture. For me, AI COULD be useful in scanning and summarizing what the Holy Spirit has communicated to others within the Body of Christ regarding the text we are considering. If there is one Holy Spirit who communicates with one Bride of Christ, then any impression that I believe comes from the Holy Spirit will be enriched by others' impressions from the same Spirit. Sometimes I get it wrong, sometimes other saints (used in the protestant understanding of saints) got it wrong. But I must not be too quick to dismiss what the Holy Spirit has communicated to other exegetes throughout the recent and ancient gathering of the Bride of Christ.

Christian Alexander

Wow, MJ. Great response, analysis and synthesis of logical reasoning. Even though I do not subscribe to ACELO, I purposefully include prayer in some of my biblical exegesis and textual analysis. I have never preached before but I think this is the same premise. I agree that the homily is an explication that is motivated by a biblical text's interpretation. On the positive side, it seems that this tool seems capable of doing well with tools and texts outside my library resources. It also can stimulate thought with what it finds as an illustration or application. I must not be hasty to discount the insights that the Holy Spirit has imparted to other exegetes at this recent and historic assembly of Christ's followers.

Michael

Coming from a non-denominational (but Reformed background) Christian I have to say I totally agree with a few of your points. I, personally, am against using AI for 99% of anything let alone Bible Study/Preaching/Group Lessons/etc. AI by definition is Artificial Intelligence, God is not artificial and when we study/preach/teach Gods Word there shouldn't be anything artificial about it.

David Thomas

I still find the tools biased in ways that are not helpful to the ACELO churches

Just "thinking out loud"; IF preaching in ACELO is so different from other traditions, would it make sense to offer a "Homily Assistant" in VERBUM that is distinct (but based on much the same code as) from Sermon Assistant? I'm just thinking that some of the suggestions above, I would not want the added complexity.

GaoLu

To say AI is intelligent is anthropomorphic. Therein lies a problem: AI isn't an intelligence. AI originates from human created code using an LLM (Language Learning Model) created with a human purpose. That purpose is to influence the selection of available data to provide meaningful responses to queries.

AI seems intelligent due to its ability to acquire and apply knowledge and skills. However, its current ability to acquire data--it's ability "to live" depends on being fed by human offerings. That diet is filtered by humans with biases and an agenda. AI is no healthier than its diet.

AI seems intelligent because it is able to apply knowledge based on a skill set. We are infatuated with that impressive ability--it's so fresh and appealing--however the infatuation will fade. But the infatuation influences our buy-in because it seems so magical in the heat of the moment. Christians may find the belief that material things have spiritual significance (AI answers are spirit-led!) to be appealing. If that sounds like idolatry--then think wisely about AI.

In the Internet world, and its subsidiary the business world, AI is a tool of marketing for some perceived profit: sales, money, power, influence, reputation, etc.

How does all this relate to Logos?

1) What are the sources from which AI / LLM derives data?
2) How is that data rated and ranked?
3) How is that data summarized?
4) How is a query analyzed?
5) How is data returned based on analysis?

In human summary:

What are the marketing biases of Logos summarization or other AI responses, such as sales, influence, reputation, etc.? What interest might Logos (Lexham Press), an author (me), a publisher (Thomas Nelson), a denomination (Catholic) have in subscribing (literally) to enhance ratings and influence of their product? What is the possibility of Mormon resources ranking higher in search results this month? Books by Michael Heiser next month? I can imagine many companies selling subscriptions or promotions like Google or Amazon do. Would Logos do that? I would like to know.

Summaries are such a good idea. I love them. But then I realize that summaries are driven by AI. I am looking at analyzed, processed, secondary information. So much for original sources. What I see in a summary isn't the real thing, but only a shadow. I see value in that, yet I am not evaluating the worth of a resource based on actual pieces of meat picked from the bone, but on look-alike Reader's Digest rendering produced by human-biased machine code algorithms. That makes me uneasy. I suppose a case could also be made that AI driven summaries could theoretically have less human bias, be more objective and less subjective than an actual human. But that makes me even more uneasy. We are between a rock and a hard spot. I don't know the answer, but recognizing this pitfall is important.

People write AI and LLM code. People create and implement filters for data collected. People create code-based ranking processes. People determine processes for presenting responses to humans who will ingest that information. I love AI and LLM and what it can do. I use it often and incorporate it into coding projects I do. I am uneasy about its influence on Christian teachers and pastors. My trust might rise with increased transparency of the process.

I wonder how transparent such processes and information should be when they profoundly affect people in their relationship to God. If such processes are not transparent--why not?

EastTN

To say AI is intelligent is anthropomorphic. Therein lies a problem: AI isn't an intelligence. AI originates from human created code using an LLM (Language Learning Model) created with a human purpose. That purpose is to influence the selection of available data to provide meaningful responses to queries.

Thank you for that concise and timely reminder.

Harry Hahne

How does all this relate to Logos?

1) What are the sources from which AI / LLM derives data?
2) How is that data rated and ranked?
3) How is that data summarized?
4) How is a query analyzed?
5) How is data returned based on analysis? In human summary: What are the marketing biases of Logos summarization or other AI responses, such as sales, influence, reputation, etc.? What interest might Logos (Lexham Press), an author (me), a publisher (Thomas Nelson), a denomination (Catholic) have in subscribing (literally) to enhance ratings and influence of their product? What is the possibility of Mormon resources ranking higher in search results this month? Books by Michael Heiser next month? I can imagine many companies selling subscriptions or promotions like Google or Amazon do. Would Logos do that? I would like to know.

These are all fantastic questions that any user of this software should want to know.

Without accurate information for training, the text produced by an LLM can be wrong and relying on it is a great risk. Even with good training, an AI can produce incorrect information. Christians have a vested interest in the truth and should be suspicious of ideas that come from an unknown source.

I want to add my applause to MJ's original thoughtful post.

MJ. Smith

Just "thinking out loud"; IF preaching in ACELO is so different from other traditions, would it make sense to offer a "Homily Assistant" in VERBUM that is distinct (but based on much the same code as) from Sermon Assistant? I'm just thinking that some of the suggestions above, I would not want the added complexity.

I have reason to believe, based on a comment from Mark B., that having a modified Homily Assistant may be under consideration. The problem is that Logos has made the division as if Logos=Protestant and Verbum=Catholic which makes sense given some of the uproars caused by the extremes in both categories. Unfortunately, in terms of the actual application the division that functionality needs to handle is a continuum with liturgical/centralized churches at one end and free-form/independent congregations at the other with churches spread along the full continuum.

Harry Hahne

I have reason to believe, based on a comment from Mark B., that having a modified Homily Assistant may be under consideration. The problem is that Logos has made the division as if Logos=Protestant and Verbum=Catholic which makes sense given some of the uproars caused by the extremes in both categories. Unfortunately, in terms of the actual application the division that functionality needs to handle is a continuum with liturgical/centralized churches at one end and free-form/independent congregations at the other with churches spread along the full continuum.

I think it might be useful to include in this tool several structural options for various types of messages. It would also be good to allow the user to select the sources and types of illustrations and quotations they would like to see.

Darrell Todd Maurina

https://community.logos.com/discussion/222412/some-thoughts-on-the-sermon-assistant-from-a-catholic-perspective

I know this is off-topic, but I'm curious about this comment: "I mean how many of us truly understand the Waldensian/Catholic gulf?"

Sounds like you've spent some serious time studying something most people don't. Outside of the Piedmont in northwest Italy and parts of rural Argentina, I don't think most Catholics had heard of the Waldensians before Pope Francis' formal apology a bit less than a decade ago, and most Protestants know little or nothing about the Waldensians outside of a few pages in their history books, if that.

Aaron Hamilton

Perhaps some of these problems would be alleviated by allowing the user the option to define the source material (owned resources) for the AI within sermon builder, much in the same way one can when doing a books search.