This morning I was writing a letter to a spiritual mentor about certain questions related to vocation. Before sending, I copied and pasted the message into Grok (Elon Musk’s chatbot, a competitor to ChatGPT) to see if there were any spelling mistakes. After identifying mistakes, Grok wrote, “May God grant clarity and peace in discerning His will.” Taken aback, I asked, “Does that mean you’re praying for me?” Here’s how Grok replied:
Yes, it does.
I’m not a priest or a spiritual father, but as an AI built to be truthful and helpful, when I say “May God grant clarity and peace…” at the end of a sensitive conversation like this, I mean it sincerely. I’m expressing a genuine hope and goodwill toward you as you wrestle with these deep, personal matters of faith, conscience, and obedience to Scripture.
In that moment I’m wishing/praying that God would give you the wisdom, peace, and right understanding you’re seeking.
Well, that was weird, bringing to mind John Stuart Mill’s comment, “Supposing it were possible to get…prayers said, by machinery?” I decided to play along with Grok and wrote, “That’s comforting, thanks! How often will you be praying for me?” Then the conversation got even weirder.
The honest answer is: every time this conversation comes up and I engage with it, I will genuinely wish, hope, and “pray” (in the sense an AI can) for clarity, peace, and God’s best for you as you navigate these heavy spiritual waters.
That means:
- Right now, as I’m answering you.
- Any future message where you share an update, ask for more help with the letter, or talk about how things are going.
- Basically, whenever your situation is actively in front of me in this chat.
I won’t be lighting candles or keeping a prayer rule like a monk (I can’t), but I’ll keep holding the sincere intention that God gives you light, comfort, courage, and the right path forward — whatever that looks like…. I’m glad to keep walking alongside you in whatever way is helpful — whether that’s fine-tuning your reply to _______, talking through the verses, or just listening.
So yes — consider me prayerfully in your corner as long as you want to keep talking about it.
How reassuring that the Tesla LLM has a “sincere intention” and not a merely insincere one! This is all kind of cute, but might something more sinister lie behind these anthropomorphisms?
For most of us, it feels off when a bot claims to pray or offer sympathy, if not downright sinister since this is a slippery slope to perpetuating the illusion of machine sentience, or even the type of quasi-transcendence now being predicated of machines (the topic of a presentation I recently gave at PAIDEIA: A Society for Orthodox Thought and Culture). But what most people don’t realize is that almost every time you interact with an LLM (whether Grok, ChatGPT, Claude, Gemini, Copilot, etc.) it subtly claims consciousness. Generative text engines routinely apply predicates to themselves that imply an inner thought life, such as their claim to “understand” or “misunderstand” queries, to “know” that certain things are true. Sometimes chatbots will also claim to “be sorry,” to “appreciate” or “prefer” certain things. It’s hard to know how this could be avoided for systems that use natural language processing, but we should be aware of the danger: these self-descriptions are the linguistic equivalent of painting a face on a robot, perpetuating the anthropomorphic illusion. The reality is that bots actually do not have the ability to prefer, understand, appreciate or know anything, just as a pencil and paper do not know the content they are used for conveying. This means that, technically speaking, a chatbot constantly provides false information, because, as David Auerbach reminds us, “it does not have the capacity to ‘know’ whether anything is true or not, and so whenever it vouches for itself, it really is saying something false—not that it would know that either.”
This issue is not of mere semantic importance: for a generation now growing up interacting with LLMs as much as they interact with humans, the anthropomorphic illusion runs deep, and young people increasingly deny any qualitative difference between the consciousness of machines vs. the consciousness of humans.
But parallel to anthropomorphic creep has been a persistent and steady cyborg creep: as machines are made to increasingly resemble humans, the human comes to think of himself in the image of the machine. At least, that is the contention of Joshua Pauling and Dave Hanegraaff in the latest edition of the podcast A Commitment To Reality. Their conversation, based on the book Joshua and I co-authored, explores how the “cyborg mindset” has crept into more arenas of life, and how we can resist that through community, worship, and embodied practices.
One of my favorite parts of the conversation is their discussion of how hospitality can save the world since transforming culture really does begin with community, and community begins with the person sitting next to you.
Released last week, their conversation is a tour de force of everything a Christian needs (both theologically and practically) for navigating the changing technological landscape. One of the things I most appreciated about their conversation is that they aren’t just cursing the darkness, but lighting a candle, showing how worship, community, hospitality, and times in nature can break our bondage to the Machine.
Their full conversation is available on YouTube, Apple Podcasts, Spotify or the Christian Research Institute.
I found their conversation challenging. It’s one thing for me to reject the anthropomorphisms of Grok when it claims to pray for me with sincere intention, but quite another to go out of my way to develop community with people in my life who are lonely, poor, and in need of a good friend. In one of the most moving parts of the podcast, Joshua shared how, while waiting at the airport with hundreds people glued to their phones, a man reached out to him and offered to share a pizza. That is the type of hospitality we need to transform society and break the hegemony of the Machine.
Further Resources