Can something written/edited by AI be considered the inspired word of a God ?

[RedTagAlan]

I am atheist, ex-born again Christian, and in a recent thread there appears to have been a fair bit of AI "tidying up" going on with a Christians posts.

I am making no accusations of course, nor commenting on the rights and wrongs, but it sure got me thinking about theists using AI for religious matters.

If a Vicar/Minister/Priest/Iman/ Theist etc uses AI, and considers that AI output to possibly be "inspired by God", and if AI can fake it, then, then do Theists believe their God can influence AI ?

It's a rather big can of worms when one thinks about it.

The question is not about God or no God, it's about if believers think AI can be divinely influenced, or inspired.

Because if they do, that's a massive can of worms. Because the people who designed the AI can explain how the AI, or LLM, constructed the text.

Now, I do not know how AI works, but it's no different to how a plane can fly. I can explain how a plane can fly, others can explain how AI works.

So for anyone to say AI produced text can be inspired, is really no different than to point at a plane, and say it is God keeping it in the air.

See what I am getting at ?

@RedTagAlan .
I think it's not that AI itself can be inspired by God, but that there is a (small) possibility on occasion that an individual could use it in a very particular inspired way.

I can see the logic in that. If the input is inspired, then the output can possibly also be considered the same.

It could be tested too. Using the same AI system, set a task to an ordained clergy and an atheist. Create a sermon on a specific subject, then do a blind panel of Christians to score what sermon they think is most inspired.

One thing for sure though. I do think if AI is used, then it should be made clear that it was.

I think it depends on how you view religious text. For me it's a guideline because its been 'touched up' many, many times by human beings over the years, translated across many different languages etc. I'm not sure why ai would be significantly different than that in all honesty. Which is why it's something to be read cautiously and carefully and I personally wouldn't take it as verbatim.

in theory god made everything, so by that logic then ai is also a product from god as humans made it or wrote the code for it, the same as the human brain wrote the bibles and other texts etc

hope this helps :

it's just a very clever computer program built by people. Basically, chatgpt or grok etc are trained by reading billions of pages of human writing (books, articles, conversations, sermons, everything online and in libraries), learning the patterns of how words, sentences, and ideas usually connect. When you ask it to create something like a sermon draft it doesn't "think" or get inspired; it predicts one word (or piece of word) at a time, choosing what is most likely to come next based on all those examples it's seen before, kind of like an extremely fast autocomplete on your phone but on a massive scale.

@RedTagAlan
Well to me one of the confounding variables there would be in such an experiment is that it matters very much who each particular minister and each particular atheist is.

The division into two sets of atheist/ minister is obfuscated by the immeasurable difference between individuals, even within each separate set.

It's about relationship (between AI interpretation, the person's resulting sermon, and the person's wisdom in relating to their congregation/audience.) The result will be unique each time.

The result will be something to do with the sermon's effect on the audience/congregation ( named according to whether atheist or minister delivers it.) For example how well the receivers remember it, or how much it has changed their thinking etc.

And of course, a sermon partly also depends on how it is delivered by its creator; yet another very important variable to influence listeners. A sermon is not only its content, but also its delivery.

These are just some of the issues that I can think of offhand that are relevant to the confines of an empirical experiment.
There must be many more.

Oh, another one would be exactly how much AI has been used in the sermon's construction eg. all of it, one sentence, or what?

I love it. At work this week I asked for a schedule of the cost of a certain product in a range of countries in local currencies and converted into one currency for comparison. I asked it to provide sources of the data and make a presentation - took less than 5 minutes. Ready for me to talk through with our client.

If you believe in an interventionist God, you believe in an interventionist God. It seems no stranger to me that God might influence the output of an AI than that he might (actively) inspire a person, part a sea, heal the sick etc etc.

OP, you might enjoy the book God Human Animal Machine by Meghan O'Gieblyn.

Yup, I have a general understanding of how AI works.

In fact, with the current thinking of AI degradation, that is that when AI output is fed through AI again and again, with no original input, then the results will lose accuracy, we might end up with totally new Bibles.

It would have to be repeated of course. And data on beliefs collected too. Calvinists are going to have different thoughts on "inspiration" than Anglicans I expect.

It's such an interesting project!
I think controls would be crazy to consider though!

It would be no different I suspect than the "power of prayer" studies ?

Sample here.

Prayer and healing: A medical and scientific perspective on randomized controlled trials - PMC (nih.gov)

No difference for clinical stuff, but potential placebo effects for stress situations ?

I would maybe do it with an AI sermon being derived by the same priest. Do an exit poll on wether it was inspired or not. But tell some groups it is AI before, and others not. See if there is any difference. But that might be more of a preconceptions result that is learned ?

Yes, I see what you're saying regarding the results of various attempted experiments.

It's the discussion that really interests me though, and I agree with the researchers when they conclude that trying to gather empirical data from any concept involving prayer is impossible.

I think any answer to prayer, or none, is all to do with relationship. Since every relationship is unique, quantifiable outcomes can never be obtained.

I think a false belief that a LLM could be channeling the word of God has been involved in several cases of AI induced psychosis, hasn't it?

Has it ? I will have a look.

Sorry @MyThreeWords I missed your post.

If I read this correctly, I think you are saying that imperfect AI is obsolete for a minister constructing a sermon, as they have the perfect tool to consult in the form of the Bible?

I think that would almost always be true.

I also think that there's just that rare occasion when AI could communicate divine inspiration for a specific minister's particular sermon.

I think the Bible talks of God revealing himself in the most unlikely situation or the most unlikely person. So why not AI?

Hi, @Myoldbear I happened to be on MN when your reply came through (I don't spend my whole life here, honest!

No, I didn't mean that AI is obsolete for a priest. I just meant that to the extent there are flaws in AI outputs (that are carried over from the flawed natures of the millions of souls that the AI is trained on), that doesn't matter, because all of those flaws represent the same type of white noise that the priest will encounter when he consults his own heart, or when he reads the bible, or when he follows any strategy whatsoever dedicated to trying to discern the word of god.

So using AI wouldn't be different from using any of these sources. So the priest should be fine with using AI - so long as he is practising the same critical discipline (the same careful. discernment) as he always must.

In your earlier reply you imply a contrast between bluebells and AI -- with the former, not the latter, being the work of god.

Surely, though, god made AI as much as he made bluebells? Sure, human activity intervened, constructing the software and the training input. But is that not just the same as the intervention of bees, whose pollination activity over the millennia has shaped the evolution of bluebells? He starts things off, and we run with it!

(I'm not actually sure if bluebells are bee-pollinated, but if they aren't, just substitute another flower, or another pollinater)

@MyThreeWords
Oh I had completely the wrong end of the stick there!
I'm going to consider a reply when I've got a bit more time.
However, I noticed you're on that pedantry thread about a poster using the expression 'hair's breadth' and it was autocorrected to hare's breath.
I think that's beautiful and I could easily see that if I were a minister/priest creating a sermon, I may see an AI intervention like that as a tiny inspiration.

What a very lovely idea. "Out of the mouths of babes and AI ..."

I've definitely read of more than one case, and it seems there is plenty of discussion about it on a quick google. So I think I would exercise extreme caution with using it in this way having read the stories.

I like what you've said in paragraphs two and three, and yes, I think you're right.

It's very like what I wanted to say, but you've expressed it more plainly.

Reading it set out as you have, I think it even looks like common sense, whereas I was sorting of making it look deeper than it is!

Now, before discussing bluebells, I should be clear that I don't think that AI is devoid of the possibility of being used in a divine way, which is what I think you felt I was saying.

I do think that the priest/minister who is working on a sermon may have more 'white noise' to overcome in discerning a message using AI compared to when they are inspired by some other, possibly more usual illumination of the will of God (eg the bible, a sunset, word of God/good in their head.)

Regarding bluebells now:
I do think that they are a more direct communication from nature/God/Good compared to AI.

This is because they just are what they are.
Yes, they are living creations so they will change, be affected by bees as you mention,(which are also behaving according to nature of course), and any other environmental variables.

However, the human input into AI output is more complex, I think.
This is in large part due to the greater choices we have in how to fulfil our human responsibilities ( eg for how we create AI input) compared to how a bluebell is purely responsible for being a bluebell.

I think it might be something to do with Free Will in how sensitively collective humanity constructs complex AI input, compared to how bluebells just fulfil their God -given, (as I would say!) nature as they must, without needing to consider if they should try to be a good bluebell or not.

I use AI a lot in my profession and I think you have misinterpreted how AI can be used. If it’s used to help create a speech, a sermon, a religious leaflet if it’s used to help understand a text and the meaning behind it. I don’t think there is anything ethically wrong with that- it’s the same as a teacher producing resources and lessons using AI tools.

but using AI to rewrite the Bible, to alter quotations from the Bible - that’s when it becomes a ethical issue

Hi @lovemetomybones ,
I don't know if you're talking to me.

I agree that there's nothing ethically wrong with AI.
What I do think is that a minister using it to create a sermon may want to be more thoughtful in how they use it, compared to someone who hasn't taken on the responsibility of guiding others along a certain path.

I use AI all the time at work - the outputs form a part of my recommendations to clients as to what they should do with their business. No different from a religious sermon ?

At work we refer to AI as our best friend.