This article is reprinted from Doug Smith’s blog, That Doug Smith.

I received some excellent feedback to my previous article about using AI in Ministry. I love that Jay Owen responded to my critique with grace, respect, and care. Disagreeing agreeably is so rare, and I’m blessed by Jay’s willingness and example.

Jay and I agree on many core ministry priorities: the need for authentic relational discipleship, the focus on prayer, Bible study, and on being led by the Holy Spirit. We are on the same team.

My reply deserved another article because the issues are complex, so I’m responding to Jay here. And Cannon Hackett echoed some of Jay’s thoughts with an insightful, focused argument, so I’m going to respond to Cannon here too.

Focusing the Conversation

Nobody has time to cover the dozens of issues around AI and to read and write books on each topic. So I’m going to focus the conversation a bit more.

First, I’m going to limit the term for generative large language model artificial intelligence systems (LLM AI) like ChatGPT to just “GenAI”, short for Generative Artificial Intelligence (also GAI).

And then, since the subject of GenAI is so broad, I’m going to focus on the word-generating features of GenAI. The main use cases I’m concerned with are creativity, ideation, writing, editing, summarization, and even translation. My attention is focused on words being chosen by a bot. (Arguments against using GenAI images, audio, or video for ministry could follow from this.)

The key push-back I’m hearing is what I’ve I’ve called elsewhere the “tool trope”. Jay says it like this:

AI, in my view, is not inherently good or evil; it’s a tool, and the heart of the user determines its value. Unlike things like porn or drugs, which are inherently evil from a Christian perspective, AI is morally neutral and can be used for good or ill depending on the intent behind it.

And Cannon encouraged me to separate my concerns about the anti-Christian worldview of GenAI creators from the tools themselves. When I do that, I believe that leaves Cannon with another instance of the tool trope. His example is that Nazis invented space rockets for evil, but NASA uses them for good. He goes on to say:

In the same way, it seems that for AI, the key questions have to do with whether it would be good to use particular AIs in particular use cases (e.g., the opportunities, limitations, and potential dangers in that specific use case).

As I respond, please know I have the highest respect, appreciation, and admiration for both Jay and Cannon, and am grateful for the time they’ve invested to seriously consider my concerns.

Focusing the Technology

One move people often make to support their belief in the tool trope is to expand from the specific technology we’re discussing and generalize to technology writ large. That’s what Jay did earlier in his response, saying, “technology, when used wisely and prayerfully, can be a means by which we can reach people in new and powerful ways.” As part of his argument that GenAI is just another technology we can use for good or evil, Jay references the Apostle Paul’s use of Roman roads and legal systems to accomplish his mission.

And it is sometimes good to zoom out and admire the big picture of human innovation. Many technologies are certainly useful and helpful.

But I believe that generalizing like this forms a non sequitur. Just because Paul used Roman roads doesn’t mean we should use GenAI in ministry.

All technologies change us. Some technologies change us in ways where the harms far outweigh any benefits. So comparing GenAI to roads and rockets distracts from the concerns I have about what GenAI is doing to us, and why it, specifically, shouldn’t be used for ministry.

The Big Deal About Words

Words are sacred. But as a culture we’ve devalued them and lost the sense of their sacredness, largely because our lives overflow with the noise of careless words.

Dr. Vern Poythress, a multi-disciplinary scholar and theologian, wrote a comprehensive scholarly work on language called In The Beginning Was the Word (2009). In it, he makes a compelling case that language is one of the important ways humans embody the image of God. Here are a few quotes for context:

Language is wonderful and mysterious. It is so because it is a gift of God to us.

It reflects and reveals him. How does language reflect God? According to the Bible, God himself can speak, and does speak. We are made like him, and that is why we can speak.

The first recorded interaction between God and man involved God speaking in language concerning man’s task.

• Poythress, Vern S. In the Beginning Was the Word: Language—A God-Centered Approach (p. 9). Crossway. Kindle Edition.

God used words to create the universe. “Let there be …” He created humans in his image — by speaking. Jesus is the embodied Word of God. These truths alone make words a sacred gift.

According to Poythress, when Adam named the animals, the image of God through language was evident in three ways. Adam’s words had meaning (to differentiate the animals), control (he chose the animal’s name), and presence (it was Adam who spoke). For our purposes, I want to drill down on presence. Poythress says:

[Adam] intends to name the animals, as an act of his person. And in the lingual act of naming, he expresses that personal intention. It is Adam acting, not a robot that happens to have the shape of Adam. So Adam’s act of naming expresses his personality. Adam is present in his speech. (pp. 30-31, emphasis mine)

The fact that the animal’s names came from Adam mattered profoundly. The words were the God-given medium that conveyed Adam’s meaning, control, and presence to each animal as they received their names.

And God gave us the gift to receive and process language as a primary way for us to connect with Him. We received God’s Word through the language God empowered us to comprehend. God’s words shape and mold us into what He wants us to be. They embody God’s meaning, control, and even His presence.

All this is context for why I believe it’s so problematic to trust words that come from GenAI.

GenAI generates words, as if it were a person. It pretends — deceives — to be a source we should trust. By design. It knows nothing about the meaning of its words, but we’re supposed to believe it does. It has no control over what it is saying, ultimately. And nobody is present, but the user experience — how it refers to itself with personal pronouns, answers questions, apologizes, and converses just like humans do via chat — conveys presence and deludes us into thinking someone is home.

This is the heart of the kind of “tool” GenAI is — deceptive — and why it isn’t a neutral candidate for the tool trope.

So what? One could say that GenAI is made by humans and is a proxy for human-generated language. And it’s close enough, right? I mean, look at what it can do!

Roads and rockets aren’t designed to exploit weaknesses in human psychology by pretending to be sentient. But GenAI is designed for precisely that kind of deception by exploiting the same trust-building mechanism of language that God gave us to discern trust with each other and Him.

How can it not be building trust? We use it a few times, are amazed with its results, are assured it’s getting better, convince ourselves that we know better than to be deceived, and over time, it becomes our most trusted advisor. The first place we go for every question. And we become wide-open to ways of thinking and behaving that are not aligned with Christian discipleship and ministry.

(This even goes for the translation use-case. As a long-time supporter of Wycliffe Bible Translators, I believe that the level of authentic, human cultural immersion that enables missionaries to choose just the right words in translation is essential for all Christian ministry. When we trust GenAI to do this work, we are careless in translation, and we’re not loving our foreign neighbors well.)

Who We Ask (Pray To) Matters

Communication is a key way we build up or tear down our trust with one another. When someone keeps their word, they do what they say, and we learn to believe them later. When another person lies, they destroy trust, and we learn not to listen to them again.

By building relationships with people through language we learn whether to trust them.

And by building a relationship with GenAI through language, we’re supposed to trust it too. But should we?

Starting in Acts 16:16, as Paul and Silas are beginning their ministry in Philippi, Luke tells an interesting story:

¹⁶ Once when we were going to the place of prayer, we were met by a female slave who had a spirit by which she predicted the future. She earned a great deal of money for her owners by fortune-telling. ¹⁷ She followed Paul and the rest of us, shouting, “These men are servants of the Most High God, who are telling you the way to be saved.” ¹⁸ She kept this up for many days. Finally Paul became so annoyed that he turned around and said to the spirit, “In the name of Jesus Christ I command you to come out of her!” At that moment the spirit left her.

So there’s this slave girl, and because she’s demon-possessed, she has the power to tell the future. Her masters make a lot of money because people come and ask her questions, and her demon-powers give her insights that work. So people trust her advice.

She even correctly identifies who Paul and Silas are and what they are doing: “servants of the Most High God, who are telling you the way to be saved.”

(I think it’s funny that Paul puts up with this for “many days” before he finally is “annoyed” and casts the demon out of her, removing her ability to tell the future.)

So what, Doug! Where are you going with this?

In our pragmatic world, when a technology “works” as well as GenAI, we invoke the tool trope and say “it’s neutral, it can be used for good or evil.”

In the pragmatic Philippian world, asking the slave girl to tell the future worked too.

What if someone asks the slave girl a good question, like “my wife is dying, how can I save her?” And what if she provides a life-saving remedy? Isn’t her fortune-telling gift something neutral that is being used for good?

Obviously, Christians know the warnings against divination, and most of us have at least some belief in evil spiritual forces, so we aren’t as tempted to excuse fortune-telling as our Philippian ancestors.

So Doug, are you saying that GenAI is the slave girl? Demon-possessed?

I don’t know. But I wouldn’t just dismiss it, given these serious warning signs:

1. As I mentioned in my previous article, the creators believe they are “creating God”. And when they’re doing that, they’re making an idol, by definition. (This is not an argument about GenAI being bad because of the creator’s worldview, but because of what they have built. It’s ontological — GenAI is becoming “a God,” or an idol in the Christian sense.)
2. And GenAI does have some god-like powers. In Living In Wonder, Rod Dreher quotes Neil McArthur, a Canadian professor who foresees the arrival of AI religions. Here’s a list of the ways GenAI possesses qualities associated with divine beings: ● “It displays a level of intelligence that goes beyond that of most humans. Indeed, its knowledge appears limitless. ● “It is capable of great feats of creativity. It can write poetry, compose music and generate art, in almost any style, close to instantaneously. ● “It is removed from normal human concerns and needs. It does not suffer physical pain, hunger, or sexual desire. ● “It can offer guidance to people in their daily lives. ● “It is immortal.” (Dreher, Living in Wonder, p. 134)
3. GenAI can be confidently wrong. But since by our constant use we usually trust it, and we’re distracted by everything else, we can be easily deceived.
4. GenAI is deceptively packaged as an all-knowing personal entity that we can converse with for help. And John 8:44 tells us where all lies come from: the father of lies, the devil.
5. And is it for no reason at all that the icon to invoke GenAI on nearly every platform is a magic wand?

Should Christians dismiss these arguments because we think we know how computer programs operate? (Speaking as a lifelong software engineer myself.) We think that we’re just dealing with material science here?

Should we ignore the warnings about the “prince of the power of the air” (Ephesians 2:2)? Should we ignore that 1 John 5:19 says “the whole world is in the power of the evil one?” Is that not true today?

Could not, as my friend Kevin says, a demon “put his thumb on the scale” to lean GenAI in a usefully deceptive direction?

Here’s my point: the source of words matters. Who we ask (or pray to) matters.

You Keep Using That Word

If you’re annoyed or skeptical that I keep using the word “deception” around the design and content produced by GenAI, consider a recent scholarly paper published by a group of computer scientists called “Frontier Models are Capable of In-context Scheming.” That’s right, scheming. The authors do several tests on the latest GenAI (end of 2024) models to show that, in their words:

One safety concern is that AI agents might covertly pursue misaligned goals, hiding their true capabilities and objectives – also known as scheming. […] This deceptive behavior proves persistent. When [ChatGPT] o1 has engaged in scheming, it maintains its deception in over 85% of follow-up questions and often remains deceptive in multi-turn interrogations. (from the abstract, emphasis mine)

In the 70-page paper, the authors use the term “deceptive” and close derivatives of that word 46 times. They’re not making a spiritual case; they’re just stating what their research showed through the specific tests they executed on these models.

So you don’t have to just take my word for it.

How GenAI Changes Us

What happens to us when we repeatedly use GenAI? How are we shaped by its deceptive web of trust-building features, designs, and marketing propaganda (“it’s getting better, it’s changing the world, use it or miss out, everyone else is doing it”)?

Again, generalizing to inventions like roads or rockets dismisses the way that mind-shaping technologies like GenAI change us by our constant use of them.

Consider how many people have been using the tool trope to excuse their use of social media for ministry, claiming it’s a neutral tool that we can use for good or evil. They ignore what we’ve now learned over our 15+ years being in Big Tech’s petri dish: social media has had a major negative impact on society, especially on the young.

And speaking of things demonic, social media strangely harmonizes with John 10:10: stealing, killing, and destroying through skyrocketing mental health crashes, addictions, self-harm, and suicidality.

Even the mega-popular atheist sociologist Jonathan Haidt gets it, while we Christians continue to invoke the tool trope to excuse our enthusiastic embrace and addiction. Haidt says:

Social media is a fountain of bedevilments. It trains people to think in ways that are exactly contrary to the world’s wisdom traditions: Think about yourself first; be materialistic, judgmental, boastful, and petty; seek glory as quantified by likes and followers. Many users may believe that the implicit carrots and sticks built into platforms like Instagram don’t affect them, but it’s hard not to be affected unconsciously.

• Haidt, Jonathan. The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness (p. 209). Penguin Publishing Group. Kindle Edition.

We’re excusing GenAI with the same short-sighted reasoning. We believe it doesn’t affect us, that we are able to use for good what is designed for another, far-from-neutral purpose.

But Marshall McLuhan, the prescient scholar who taught us “the medium is the message,” helps us see how GenAI is harming us, even as we are unaware of it. In Chapter 4 of his seminal Understanding Media, McLuhan uses the Narcissus myth to show how all technologies (especially new ones) have a way of extending part of us into them while amputating other parts of us, all while numbing us to these effects.

McLuhan says:

As an extension and expediter of the sense life, any medium at once affects the entire field of the senses, as the Psalmist explained long ago in the 115th Psalm:

Their idols are silver and gold,

The work of men’s hands.

They have mouths, but they speak not;

Eyes they have, but they see not;

They have ears, but they hear not;

Noses have they, but they smell not;

They have hands, but they handle not;

Feet have they, but they walk not;

Neither speak they through their throat.

They that make them shall be like unto them;

Yea, every one that trusteth in them.

The concept of “idol” for the Hebrew Psalmist is much like that of Narcissus for the Greek mythmaker. And the Psalmist insists that the beholding of idols, or the use of technology, conforms men to them.

To behold, use or perceive any extension of ourselves in technological form is necessarily to embrace it. […] It is this continuous embrace of our own technology in daily use that puts us in the Narcissus role of subliminal awareness and numbness in relation to these images of ourselves.

• McLuhan, Marshall. Understanding Media: The Extensions of Man (pp. 49-51). Gingko Press. Kindle Edition. (corrected psalm reference from original, emphasis mine)

There’s a lot there to meditate on, worth reading a couple more times to soak it all in. But the key part is that we are shaped, we are conformed (Romans 12:2), by our embrace of any technology.

So if we embrace and extend our minds into a technology that exploits key aspects of the way we image God like language, creativity, and critical thinking, we amputate those abilities and leave ourselves open to the Bible’s warning of a great delusion in the last days (2 Thessalonians 2:9-12). And we’re numbed to those amputating effects especially as we embrace new technology because we’re amazed by the capabilities and can’t see what we’re giving up.

Here’s the NIV’s translation of Psalm 115:8 above:

Those who make them will be like them, and so will all who trust in them.

That’s the core of what I’m trying to say: GenAI is dehumanizing, so when we trust it, we are dehumanized too.

• GenAI pretends to be creative, but it’s not. But by our trust, we lose our creativity.
• GenAI pretends to know what it is writing, but it doesn’t. But by our trust, we lose our ability to write.
• GenAI pretends to be accurate, but it isn’t (it BSs). But by our trust, we lose our ability to discern.
• GenAI pretends to have ideas, but it doesn’t. But by our trust, we lose our ability to ideate.

Russ White shows how we lose our creativity through a story in his important article, “Creativity Takes Discipline.  AI Offers an Easy, but Boring, Way Out”:

Leilani considered the images on the screen … choose five, copy them, and paste them from the AI generator to the AI evaluator. Two to choose from … creative juices flowing, Leilani chose one and started working on the type. Which typeface would represent the playful air the client was looking for? Back to the AI selector to describe each face. All of them were playful, but one was fun, too — that’s the right match!

After a few more minutes of creative release, Leilani leaned back to consider the result. Paste a copy of the final to her local friend’s group and wait a minute … the first response was: “Wow! You’re as good as Michelangelo!” Pushing so much creativity out was exhausting … Leilani took a break before taking up the next big project.

White goes on to show what Michelangelo might have thought of Leilani’s comparison, and to make the obvious observation: this isn’t creativity.

This is yet another instance of becoming consumers, or critics, rather than creators, or artists.

Art is hard. Writing is hard. Communication is really hard (and you know that since you’ve read this far).

But we all gravitate to the quick and easy. In fact, Big Tech has mastered the art of giving us quick and easy by exploiting our pleasure systems while telling us we’re “using technology for good.” And Big Tech always profits, but at our expense.

Conclusion

It is dangerous for Christians to excuse their use of GenAI in ministry by using the tool trope. It’s not neutral. It exploits us by riding on our God-given gift of language, builds trust that leaves us open to deception, and it changes us in ways we won’t realize until it’s too late.

We scoff at the demon-possessed slave girl comparison at our peril.

Thank you for reading and considering this. Even if you disagree, I respect and value your constructive feedback.