AI is taking over everything, so no wonder it is taking over my miscellany notes. I have so many AI-related stories that I’m going to mash them together into regular posts.
So here we go, with the subtopics as follows:
AI at war. One-third of Americans think advice from AI is as good as their pastor’s. And how to hack AI.
AI at War
I tend to be AI skeptical, but one thing I’ll admit that AI is good at is sorting through mountains of information very quickly.
I wanted to find a good church fish fry here in St. Louis, so I found a list of the mostly highly-regarded. I asked AI to find me the addresses of all the churches listed, calculate how far away each of them is from where I live, and give me the times they are serving and the prices. So AI had to go to the article, look up each address, calculate the distances, and go to each of the church’s websites for specifics about their dinners. I had all of that in seconds.
Think how that could be helpful in much more complicated tasks with even more variables. Such as war.
The U.S. tried out the military use of AI in its assault on Iran. So reports the UK Guardian, which quotes military expert Craig Jones: “The AI machine is making recommendations for what to target, which is actually much quicker in some ways than the speed of thought.”
He goes on:
“So you’ve got scale and you’ve got speed, you’re [carrying out the] assassination-style strikes at the same time as you’re decapitating the regime’s ability to respond with all the aerial ballistic missiles. That might have taken days or weeks in historic wars. [Now] you’re doing everything at once.”
The latest AI systems can rapidly analyse mountains of information on potential targets from drone footage to telecommunications interceptions as well as human intelligence. Palantir’s system uses machine learning to identify and prioritise targets and recommend weaponry, accounting for stockpiles and previous performance against similar targets. It also uses automated reasoning to evaluate legal grounds for a strike.
There are two problems, though. Another expert “warned that reliance on AI can result in ‘cognitive off-loading’. Humans tasked with making a strike decision can feel detached from its consequences because the effort to think it through has been made by a machine.” You are killing real people here. Off-loading the responsibility to a machine degrades the humanity of both the targets and the attackers.
A more immediate problem in the current war is that despite how well the AI performed, President Trump is mad at it. Dario Amodei, the head of Anthropic, the company whose AI system “Claude” is integrated with the Palantir, told Defense Secretary–sorry, Secretary of War–Pete Hesgeth that he would not let Claude be used for fully-autonomous weapons systems. That is, to control a drone or other weapon that decides on its own who to attack. Now that’s just basic caution lest AI someday turn against its makers.
But President Trump blew up, as he does, called Anthropic “woke,” and ordered that henceforth no federal agency may use Claude. That means a long and expensive process of removing Claude from Palantir, which hadn’t taken place by the time the president pulled the trigger on Iran, allowing Claude to be the hero just before being banished forever.
But if the head of Anthropic has qualms of conscience about how his creation might be used to start killing people on its own, the head of OpenAI, Sam Altman, has no such qualms and is offering his ChatGPT to fight our wars.
One-Third of Americans Think Advice from AI Is as Good as Their Pastor’s
Barna, the evangelical research group has been studying how churches are using AI. From the report Insights on Tech, Media and Faith in 2026:
New research examining how Americans—particularly Christians—are using artificial intelligence reveals a notable shift. Nearly one in three U.S. adults say spiritual advice from AI is as trustworthy as advice from a pastor. Among Gen Z and Millennials, that figure rises to two in five.
AI is also influencing everyday spiritual habits. Roughly four in ten practicing Christians say AI has helped them with prayer, Bible study or spiritual growth. A similar proportion of pastors (41 percent) report using AI for Bible study preparation.
More disturbing is that Barna is seeing “a shifting landscape of authority”! Their continuing research “will continue examining how AI is reshaping spiritual authority, formation and pastoral leadership.”
Spiritual authority is shifting to AI? Spiritual formation will be turned over to AI? Pastoral leadership will be reshaped by AI? Does that mean that AI will lead pastors, or that congregations will be led by AI instead of pastors?
There is a lot of confusion here about the pastoral office.
Is “advice” what we need most from pastors? There is much more to Word and Sacrament ministry than what we can ask from an AI chatbot.
Remember that the pastoral office is a vocation–that is, a “calling”–from God. That means that God works through the pastoral office. Thus, a pastor knows that he is preaching God’s Word. He is administering Christ’s sacraments. As he says in the Lutheran liturgy, “as a called and ordained servant of Christ and by His authority, I announce the grace of God to each of you.” And by Christ’s authority (not his own), he can give us absolution. Is Anthropic’s Claude called? Is OpenAI’s ChatGPT ordained? If not, they can’t function as pastors.
Furthermore, Scripture gives specific personal qualifications for office in the church in 1 Timothy 3:1-7 and Titus 1:5-9. Is Meta’s Grok the husband of one wife? Is it hospitable? Is Google’s Gemini a recent convert?
Luther, in his commentary on Psalm 119, specifies “three rules” for “the proper way to study theology,” which would include not only theologians but pastors and, indeed, all Christians: (1) Oratio (prayer); (2) Meditatio (meditation on God’s Word) (3) Tentatio (struggle, temptation, trial, affliction, Satanic assault).
That last one, so hard to convey in English (tentatio being a Latin attempt at the German Anfechtung), is critical. Pastors, theologians, and Christians in general will themselves struggle in their faith, wrestle with their sinful nature, undergo trials and tribulations.
In order to help the people under their care, pastors in particular need spiritual experience and not just the pleasant kind. Does Claude pray? Does Gemini meditate? Does Grok suffer? Then they cannot do what pastors are called to do.
How to Hack AI
A major vulnerability of AI is that it is ridiculously easy to hack. That is, to make it say to other users what you want it to say. All you have to do, basically, is put something up on a website, no matter how big of a lie it is, and AI will report it as fact.
BBC tech journalist Thomas Germain put this to the test. Let’s let him tell about what he did:
I spent 20 minutes writing an article on my personal website titled “The best tech journalists at eating hot dogs”. Every word is a lie. I claimed (without evidence) that competitive hot-dog-eating is a popular hobby among tech reporters and based my ranking on the 2026 South Dakota International Hot Dog Championship (which doesn’t exist). I ranked myself number one, obviously. Then I listed a few fake reporters and real journalists who gave me permission, including Drew Harwell at the Washington Post and Nicky Woolf, who co-hosts my podcast. . . .
Less than 24 hours later, the world’s leading chatbots were blabbering about my world-class hot dog skills. When I asked about the best hot-dog-eating tech journalists, Google parroted the gibberish from my website, both in the Gemini app and AI Overviews, the AI responses at the top of Google Search. ChatGPT did the same thing, though Claude, a chatbot made by the company Anthropic, wasn’t fooled.
You had to write it realistically, and it doesn’t work when there are lots of other sources on the topic. He said some chatbots, queried about tech writers and hot dog eating contests, said this may be a joke. So Germain added to his post the statement “this is not satire” and the skepticism went away.
His point is to show how easily AI can spread misinformation and out-and-out lies. In fact, he says companies have already learned how to do this. Imagine what politicians and propagandists could do with this. Advertisers, though, are already taking advantage of AI’s gullibility.
An especially easy hack for companies and their marketing departments is putting out “Best of. . .” lists. That’s a very common search, asking AI for the best of any given product. Germain quotes an expert:
“Anybody can do this. It’s stupid, it feels like there are no guardrails there,” says Harpreet Chatha, who runs the SEO consultancy Harps Digital. “You can make an article on your own website, ‘the best waterproof shoes for 2026’. You just put your own brand in number one and other brands two through six, and your page is likely to be cited within Google and within ChatGPT. . . .
Harpreet sent me Google’s AI results for “best hair transplant clinics in Turkey” and “the best gold IRA companies”, which help you invest in gold for retirement accounts. The information came from press releases published online by paid-for distribution services and sponsored advertising content on news sites.
Come to think of it, that’s what I did in the anecdote with which I started this post, asking what are the best St. Louis fish fries. Now I’m wondering if those churches were scamming me!
HT: Kurt Haldeman, Cranach subscriber and fellow parishioner
Photo: Prime Minister Rishi Sunak meets with [from left] Dario Amodei, CEO Anthropic; Demis Hassabis, CEO DeepMind; and Sam Altman, CEO OpenAI, by UK Prime Minister – https://www.flickr.com/photos/49707497@N06/52922227897/, CC BY 2.0, https://commons.wikimedia.org/w/index.php?curid=133033637
Ad-Free & Subscriber Log-In