In the last post I borrowed a story from John Lennox — the deepfake channel built in his name, the fabricated lectures his own audience believed — to sharpen a question I keep returning to: is your voice still yours? Since then I’ve sat with the book that story comes from, God, AI, and the End of History, and I want to give it a post of its own. Not a review — a witness statement. Because what Lennox provides, more rigorously than anyone else I’ve read, is the theological ground under the thing I’ve been building.
A word on why he’s worth your attention even if theology isn’t your shelf. Lennox is an Oxford mathematician who spent a career in the hardest of hard sciences and a parallel career as a public Christian voice — debating Dawkins and Hitchens, writing carefully about science and faith for forty years. He is living proof that you don’t have to choose between intellectual rigour and religious conviction, and that matters here, because the question of what AI may and may not do to a human life is exactly the kind of question that gets answered badly by people who have only one of the two.
The doctrine of use
The practical heart of the book is what I’d call a doctrine of use, and it’s almost disappointingly sensible. Use AI, Lennox says, the way you’d use a Bible dictionary: to find things out, to check facts, to draft a letter — which you then edit carefully. The tool extends your reach; it must never replace your judgment. What he adds, and what most AI commentary misses entirely, is the test that tells you which side of the line you’re on. I think of it as the custodial test: does what goes out under your name still carry your responsibility?
Sit with that one. Not did a human review it — the compliance version. Not is it accurate — the quality version. But: when this lands in someone’s inbox, someone’s feed, someone’s life, is there a person standing behind it who can say this is mine, I answer for it? The deepfake channel is what the world looks like when that test fails — words going out under a man’s name and face that carry no one’s responsibility at all.
Readers of this series will recognise the shape. The Faithfulness Coach exists to ask am I being faithful to what I said matters most? — and faithfulness presupposes an I who answers. The custodial test is the same conviction stated negatively: there are things that cannot be delegated, and accountability for your own life is first among them.
Why judgment can’t be delegated — the actual argument
Here’s where Lennox goes deeper than the encyclical I wrote about two posts ago, and deeper than I had gone myself. It’s one thing to prefer keeping humans in the loop. It’s another to say why the loop has a human-shaped hole in it that nothing else can fill.
His answer is as old as Genesis: the imago Dei. Human beings are made in the image of God — and whatever else that means, it means consciousness, moral agency, and responsibility are not features that emerged from computation and might therefore be reproduced by it. AI simulates intelligence, often brilliantly. It does not think, feel, see, or know what it is like to be anything. It can produce the outputs of judgment without ever exercising judgment, the way a recording produces music without ever hearing it.
You may not share the premise. But notice what the argument does: it gives a reason why authorship is non-transferable, where most AI-ethics writing simply asserts it. If responsibility is constitutive of what a human being is — not a function we perform but part of what we are — then handing it to a machine isn’t efficient delegation. It’s self-amputation. That is the theological floor under everything I’ve written in this series, and I’d rather name my floor than pretend I’m standing on air.
Three witnesses, one conclusion
What strikes me most, three posts into writing about this, is the convergence. A Catholic encyclical drawing on a century of social teaching. An evangelical Oxford mathematician arguing from Genesis and from his own deepfake. And — far more modestly — a Finnish-German consultant who built a system to run his own life and discovered, in practice, that the moment the system starts deciding instead of asking, it stops serving him. Different traditions, different methods, same verdict: a tool, never a master.
In my consulting work I’d call that triangulation, and I’d tell a client to take it seriously. When independent lines of reasoning from independent starting points land on the same conclusion, the conclusion is probably load-bearing.
And to be honest about the book’s own centre of gravity: Lennox reads the Book of Revelation alongside the futurists, and some readers will come for exactly that. I’m deliberately not building on it here — partly because speculative prophecy mapping is a game Lennox himself warns against, and partly because the book’s real point is simpler and sturdier: the future is not in the hands of the machines, which means we are free to use them without fear and without worship. Hope, it turns out, is a practical engineering constraint. It determines how tightly you grip your tools.
What this changes on a Tuesday morning
A witness is only useful if it changes practice. Two things have already shifted in how I run the Faithfulness Coach.
First, the custodial test has joined the daily question. Alongside does what I’m doing today bring me closer to where I want to be?, I now ask of anything the system helped produce: does this still carry my responsibility? An email an assistant drafted, a plan an AI proposed, this very post — the test is the same, and it has teeth precisely because most days the honest answer requires me to slow down and actually read the thing.
Second, Lennox prescribes what he calls electronic fasting — deliberate, periodic stepping back from tools engineered to hold your attention in the perpetual now. I recognise this immediately as a cousin of something my own system depends on: the discipline of regularly climbing out of the daily stream to a higher vantage point. That rhythm — how it works, why it’s the part of any personal system that fails first — is the next post, and I’ve owed it to you since the naming post promised it.
The same open question
The Faithfulness Coach remains what it was: a personal system, not a product. But the question I asked two posts ago stands, and Lennox has only sharpened it. If you’re building something like this — a system meant to strengthen your judgment rather than replace it — I’d like to hear from you, and I’m especially curious how you answer the custodial test in your own work. Get in touch at martin@bilboconsult.com or on LinkedIn.
This is the fourth post in an ongoing series on the Faithfulness Coach. Earlier posts: The Intentional Living Engine introduces the framework; What the Pope’s AI Encyclical Says to Anyone Trying to Live Faithfully connects it to AI and human judgment; Why I Stopped Calling It a Personal Operating Model explains the name. Reference: John Lennox, God, AI, and the End of History* (2026).*
Martin Kallenbach is an independent IT management consultant based in Helsinki. He helps organisations with transformation, service integration, and governance — and is applying the same thinking to the question of how to live well and intentionally.
Leave a Reply