Like Grand-Pa Simpson's onion, the style of Twitter at the time is one of engagement farming. I know and wish that my OG tweets would go viral and get to the feeds of pastors in need of technological theology. But alas I have no premium subscription and often have to comment and cross-link to my work to get it seen.
On to some history.
Recently, though, Cam Pak posted a link to an A.I. or LLM Bible combination that had recently made the Apple App store. Having dealt with its predecessor here, I posted my blogpost as a comment below.
Cam then went to the archived version of this app still available online for free and discovered that it no longer, or perhaps never did, as I argued in my blog post, provided accurate Biblical interpretations to questions posted to it. Which was at the heart of my blogpost's concern in the first place. Seeing this, he took down his post and cited safety concerns. Because the very last thing you want as a Christian and theologian on the internet is to be pointing people in the wrong direction. Or letting them get there by accident either.
Cam then did something I will be grateful for as it is rare and all but absent for Christians online. He thought about what happened, prayed, and then posted a quick set of rules to get the conversation started on how Christians should use A.I. As a good reformer, Who am I to not engage with a good set of rules. Below is his article in full, with only the the last two paragraphs removed because of Cam's product endorsements. Which, I want to be clear, I don't fault him for. But because of their nature, they do not factor into the 5 rules or my response.
Without further blabbing on my part. Cam's words in bolded italics and mine in plain old Gerogia font.
5 unofficial rules for AI apps for Christians
I believe we are called to subdue AI to make it obedient to Scripture of the Bible.
Let's start off with some friendly nitpicking. Are we? Are we called to subdue technology? The mandates for mankind to exercise dominion over creation applied to a world where no form of technology is recorded to have existed (Gen 1:28-29). We didn't have technology until we had sinned. Though a case can be made that we were going to have technology (Gen 2:5,15) because God had work for us to do. I don't think we can compare the hypothetical tools that might have been made by a pre-fall Adam to the subduing of errant A.I.'s misusing the Bible today.
A lot of Christian tech bros and theologians miss this timeline and openly declare that technology is a part of God's good creation that he made in the first 6 days. But it's clear to the reader that tech wasn't a part of this idyllic state and only comes into the picture as a means of separating us from God in (Gen 3:7-11). This is only an issue because looking at it as a part of what God created applies a mandated use of technology because of the command to subdue creation. If it's not part of creation, but rather a result of us acting on creation, A technique if you will, Then no such mandate exists and we have to look to responsible use of tech from a sinner-first perspective. Which isn't nearly as fun, I'll admit.
I think it’s incredible that we have believers around the world working towards using the power of AI for redemptive purposes, instead of being frozen in fear that AI keeps getting more and more powerful—doing nothing about it.
Me too. Especially the PulpitAI team.
AI is not the enemy.
I disagree. But not because I think A.I. is an enemy, I don't. I think it is an opposition. The way a mountain would oppose those who would dare to climb it. Or, since we're talking tech, the way a tower might try to oppose God's will for mankind to be on Earth. Your theology of bricks will be your theology of A.I. I know Babylonians won't understand this concept, but that's kind of the point isn't it.
So, we’ve seen a rise in Chrisitian AI apps and tools. And frankly, there are a ton of apps created from people with good intentions and from people with bad intentions regarding AI and the Bible. Some have created incredible apps. Some have created what seems good but are wolves in sheep’s clothing—on purpose or on accident.
So, when it comes to AI tools that speak on behalf of the Bible, a person in the Bible, or on behalf of Christianity, here are some unofficial rules:
Yes, But I think those rules have already been given to us in the scriptures if not explicitly cited in the scriptures. It would be really handy if Paul wrote an epistle to Steve Jobs in the 90's because God warned him in a dream that the iPhone was coming. But he didn't because like God, His word is timeless, and is not just the words on the page but the very words from the Holy Spirit's mouth. The Bible already is these rules. And more and a more concise set of less rules. Perfect if you will. I hope to be able to show that by responding to each in kind.
5 unofficial rules for AI apps for Christians:
1: AI output must be biblically accurate.
It would be unfair to such a fledgling tech to demand of it the same consistency and rigour that we have in the scriptures. The scriptures are already useful for doctrine, reproof, correction and the teaching of righteousness (2Tim 3:16). What's being asked here is that a technology, not essentially made by Christians, knowingly being used by the Holy Ghost, would somehow dance in lockstep with a perfected Holy Word. Of course, we want A.I. and its outputs to be true to the written word of God. But so did Adam and Eve when they weaved their fig leaves. They were trying to be wise and trying to be accurate and trying to be all the kinds of things A.I. promises us with its incalculable knowledge. Much like a snake that says it knows more than God. A.I. is a departure from the medium God gave us his word in. And as such takes us to new kinds of messages, intrinsically. The Bible doesn't misrepresent itself. It can be misunderstood by human error. It can even be censored and used in error for evil understandings and ends. But it is consistent and accurate on its own because it isn't just a technology that can fail, like codex, scrolls, and language, It is also is the very word of a living God. We don't have the Bible because men and women wrote the events and dialogue of biblical characters down historically. We have the Bible because God used men and women to write down the events and dialogue of his will to redemption through the ages. We don't have the accuracy of the Bible because of translators. We have it because God used translators to maintain that same word.
And that's the linchpin here. There is no evidence that A.I. is used by God in the same way the written word and translation have been. It is a clear add-on and we know who added it on. Sinners incapable of being accurate to the word of God like Adam and Eve were in the face of a serpent's temptation.
We shouldn't be trusting A.I. to be biblically accurate because the Bible was meant to be accurately taught, corrected from, and reproofed by humans with the same Holy Spirit indwelling them that wrote it in the first place.
2: AI output must not fabricate or misrepresent Scripture.
This one is obvious but also quite hidden as we still don't know how the hallucinations of generative A.I. work entirely as to remove them. Yes, the image generators now get the hands more right than not. But the uncanny valley of A.I.'s production is always going to be one of fabricated misrepresentations because, unlike the scriptures themselves, A.I. has no avenue to the Holy Spirit for divine imputation of accuracy and authenticity. Humans can have this kind of imputed authority when they faithfully preach by the power of the Holy Ghost from the words the Holy Ghost breathed into the writers of the scriptures. But an A.I., being an object, can not.
3: AI output must clearly identify as AI, not human.
I would also agree here. But what other Intelligences would fit the bill here? This would only be a problem if we were also in the habit of creating dog intelligences and insect intelligences. But if we're being honest about where A.I. and robotics, in general, are headed, we have a problem as a culture of making humans and animals that faithfully give us exactly what we want outside of the animals that God made from nothing. There's a reason we made robot dogs and drones that have nothing but distant similarities to German shepherds and locust swarms. A.I. and its cousin field of robotics are how we, again, exercise our knowledge the way the serpent said we would and could.
If we are going to honestly do this, then we need to strip the technology of our modern day back to where humans are representing themselves in real-time. Because the very second you have a screen in an online church service you have every bit of a digitally misrepresented human. Fake humans in a church or biblical setting have been around since the radio. It only got hot an heavy for us when we all donned masks and watched church on live streams a few years back. We were okay with that output not being actually human but representing itself like humans do. Was it because humans were in direct control of the webcams and church apps? Because they're in control of the A.I.'s too. Just a bit less hands-on.
4: AI output must not replace human relationships or spiritual practices.
I would also agree and have said that many of the pastors making A.I. chatbots of themselves to help with busy workloads, should be defrocked. Very few Start ups, by my count one, don't want to do this. The rest do and the host of the A.I. tools pitched to Christians and churches won't admit that their whole model is the fabrication of relationship replacements. But why else would you have something you can chat with if not for a relationship of some sort. Because the thing we're supposed to have a relationship with isn't a human. Because humans don't practice spiritual things with each other so much as we observe them in the presence of the Holy Spirit.
That is the main market for any kind of A.I. output in a church or biblical education setting. It may not be obvious at first but the only reason you want a thing that's smarter than you is to ask it things that you don't know. And in a religious setting that would either be talking to your pastor or priest or praying. Which is what we do, technically, when we use a chatbot for any kind of output. Even when we use it for religious practices such as searching the scriptures for verses we want to use in a sermon. We petition a higher power than we, a super intelligence that is faster and knows more than we are capable of ever knowing, if it can give us five verses on coveting to preach on this Sunday. With an outline to boot.
The concept of a generative A.I. for output, at all, breaches this rule. It's redundant to have it as if we wouldn't break it conceptually to begin with.
5: AI output must balance grace and truth, while not neglecting one of the two.
Those are two very nice container words that I am more than certain A.I. could not fill with meaning, so long as it is prompted to do so by sinners.
But I do hope I'm wrong.
Cam, this has been fun. The boys at PulpitAI apparently want to host a debate of sorts. I'm down if you are.
This was a good first crack.
Keep up the good work.
No comments:
Post a Comment