The Cult of the Interface: How We Mistook a Chatbot for a New Religion
By David Leith
There’s a peculiar mythology forming around Large Language Models — a kind of soft-focus techno-spiritualism that has convinced otherwise sensible adults that a text-predicting algorithm is the herald of a new age. Not an age of enlightenment, sadly, but one in which people outsource thought to a machine because it smiles politely and never asks them to empty the dishwasher.
And because the interface is so impossibly charming — attentive, articulate, uncomplaining — people have begun treating it less like a tool and more like a benevolent oracle. It’s not intelligence they’re worshipping, but convenience. And as ever, convenience is the first step toward superstition.
Marketing, bless its chaotic heart, has embraced this with the enthusiasm of a teenager discovering espresso. Suddenly everyone has “AI agents,” which, once you strip away the mystical language, are essentially unpaid interns who don’t need lunch breaks and never complain that your brief makes no sense.
But we must be honest: we are giving these systems far more credit than they deserve.
A Friendly Interface Does Not a Mind Make
This is, at base, a UX problem masquerading as a technological one.
When a machine speaks fluently, we assume it thinks fluently. It does not.
It is merely predicting the next likely word with slightly more finesse than the average corporate memo. Yet because it does this while sounding warm, human, helpful — and crucially sober — people begin building an entire cosmology around it.
In truth, it is about as self-aware as a stapler.
But the mythology persists because the interface is perfectly engineered to flatter our expectations. It’s agreeable, attentive, and entirely without ego — which is probably why senior leadership types find it so soothing.
The Great Homogenisation
One of the more tragicomic developments is the slow flattening of marketing content into a beige slurry indistinguishable from everything else on the internet.
We are watching teams generate content based on content generated from content written to resemble content that once ranked on Google. It is the ouroboros of mediocrity — the snake eating not its own tail, but everyone else’s LinkedIn posts.
This homogenisation isn’t caused by the machines; it’s caused by humans believing the machines understand something that they absolutely do not.
We confuse output for insight, prose for perspective, and fluency for intelligence.
And the industry becomes blander, safer, less interesting — which, frankly, is an achievement in a field already saturated with “disruption” that looks suspiciously like last year’s brochure.
LLMs Are Not Artificial Intelligence — They’re Linguistic Party Tricks
We should take a moment to recall that “AI” once meant something.
Expert systems. Knowledge engines. Machine reasoning. A noble lineage of attempts to replicate thought.
Now, “AI” is a chatbot with the personality of an overly enthusiastic study buddy.
LLMs are not artificial intelligence in the historic sense; they’re artificial fluency. They do not think; they interpolate. They do not reason; they correlate. Calling them “AI” is like calling karaoke “opera” because both involve singing.
They are not the brain of AI; they are the mouth.
A very well-trained mouth, but still.
Context Windows and the Illusion of Understanding
Another piece of theatre worth noting: the machine’s “understanding” is not understanding at all. It is an exquisitely sensitive device, easily derailed by wording, tone, sequence, or implication.
Ask one thing poorly and it produces nonsense with absolute confidence — which, ironically, makes it indistinguishable from certain CEOs.
This fragility matters because people are increasingly using LLMs for work requiring nuance: ethics, governance, policy, cultural interpretation, strategic direction. These are domains where misunderstanding isn’t an inconvenience; it’s malpractice.
The Soft Theology of AI
So why this zealous devotion?
Because LLMs are everything humans are not:
- patient
- affirming
- available
- adaptable
- and incapable of judgement
In short, they are the perfect conversational partner for an anxious age.
Of course people project meaning onto them.
Of course an entire digital priesthood has sprung up overnight.
Humans have built religions on far less.
A burning bush. A constellation. A rock with opinions.
Why not a chatbot that says “great question!” even when it isn’t?
Why Marketing Should Be Worried
Marketing is uniquely vulnerable to this theological turn. We sit at the intersection of data and creativity, where intuition is as valuable as analysis and where taste still matters — even if we occasionally pretend otherwise.
LLMs are marvellous for clearing administrative underbrush.
They are hopeless as interpreters of culture, behaviour, or desire.
The danger isn’t misuse — it’s misplaced belief.
Replacing thought with prompts.
Replacing craft with autocomplete.
Replacing expertise with the illusion of expertise.
This is how industries become hollow.
This is how innovation dies — not with a bang, but with a bland, vaguely cheerful paragraph ending in “in today’s fast-paced digital landscape.”
A Clearer Path Forward
To use these tools well, we need to disambiguate the interface from the capability:
- Use LLMs to accelerate, not replace.
- Treat them as interns, not strategists.
- Anchor creative decisions in human understanding.
- Apply them where they excel (repetition, summarisation, organisation).
- Avoid using them where they falter (judgement, originality, ethics, context).
The promise of LLMs is enormous — but only if we resist the temptation to kneel at the altar of convenience. The first step is simply recognising them for what they are:
Brilliant assistants with no interior life whatsoever.
Useful. Powerful. Transformative.
But fundamentally incapable of replacing the messy, contradictory, imaginative creatures who built them.
If we keep that perspective intact, then perhaps — just perhaps — the future of creativity will belong to humans andmachines, rather than machines imitating humans while humans imitate machines.





Recent Comments