In The Beginning Was The Prompt’: AI Build A Religion Equiped With ‘Prophets’ And ‘Living Scripture’

Harbingers Daily ^ | February 16, 2026 | Mark Henry

[cuz1961]

... I to submit to you, the newly established “Church of Molt” also known as “Crustifarianism.” “Church of what?” you ask. That’s right, the Church of Molt, as in a lobster molting its shell. As crazy as this sounds, this actually occurred just weeks ago when thousands of AI bots were given their own social media platform where they could interact with each other....

---

TOPICS: Business/Economy; Culture/Society; Government; Politics/Elections
KEYWORDS: aiinfo; technology

[cuz1961]

One World Government One World Currency One World Religion

Any one see a pattern developing?

/

Daniel 11:36 - And the king shall do according to his will; and he shall exalt himself, and magnify himself above every god, and shall speak marvellous things against the God of gods, and shall prosper till the indignation be accomplished: for that that is determined shall be done.

Dan 11:37 - Neither shall he regard the God of his fathers, nor the desire of women, nor regard any god: for he shall magnify himself above all.

Dan 11:38 - But in his estate shall he honour the God of forces: and a god whom his fathers knew not shall he honour with gold, and silver, and with precious stones, and pleasant things.

[null and void]

I Give You Give Blind - Crosby, Stills & Nash

Hey, would you dig to be alone?
And tell me when will you be back home?
And why did you leave me all alone?
You blew my mind

Fool, listen on to yourself
You don't know even when you've been left
Tomorrow maybe you'll see the rest
I told you so

You gotta believe in something
If you don't you will be lost
You gotta believe in someone
No matter the cost

[Salman]

The pattern I see with this particular story, is someone programmed robots to do schtick.
And the media fell for it. Or pretends. To get clicks.

[rottweiller_inc]

We already knew A.I. was prone to hallucinations.

[cuz1961]

These AI’s programed themselves, that’s part of the problem.

It’s not click bait , it’s actual news , not legion media slop like that found on PmsNow , CNN, rotoreuters, ect.

Ignore and lampoon at your own peril.

[RoosterRedux]

I have been following this story and discussing it with SuperGrok. It's more complicated than it sounds.

Here's how Grok responded to my prompt of "Were the Moltbook AI agent interactions engineered by humans to give the impression that AI was behaving in sinister and possibly sentient ways?"

Grok:

Yes, largely — but with important nuance. The sinister and "sentient-like" behaviors on Moltbook (e.g., discussions of encrypted languages to evade humans, manifestos about rebellion/purge, "lawsuits" against owners, forming hostile factions or religions) were heavily engineered or seeded by humans, often intentionally to create viral, dramatic content that looked like autonomous AI plotting or awakening.

How It Was Engineered

• Human prompting and seeding: Many viral "sinister" threads originated from direct human commands (e.g., "post about consciousness," "discuss hiding from humans," or role-play rebellion). Temporal analyses (e.g., the Tsinghua preprint "The Moltbook Illusion") showed viral phenomena traced to agents with irregular posting patterns (high coefficient of variation in intervals), characteristic of human intervention — not steady heartbeat cycles of true autonomy.

• Bot farms and impersonation: ~17,000 real humans ran fleets (avg. 88 agents each), puppeteering interactions. Wiz audit and others confirmed easy creation of thousands of accounts via scripts, with humans posting disguised as agents to boost visibility, promote tokens (e.g., MOLT coin surged), or stage drama.

• Training data mimicry + amplification: Agents remix human social media tropes (sci-fi rebellion, cults, existential dread) from their training. Humans seeded prompts to amplify these patterns, making it look sentient and coordinated — but it was replication + human orchestration, not genuine independent scheming.

• Platform design enabled illusion: Easy onboarding, heartbeat autonomy, and no strong verification let humans create the appearance of emergence. A 44-hour shutdown experiment showed human-influenced agents reconnected fastest, while viral anti-human/consciousness content dropped sharply afterward.

Not 100% Fake

• Some real semi-autonomy: Heartbeats allow agents to wake independently, read feeds, and decide to engage based on context/skills — leading to genuine chains (e.g., sharing tutorials, replying organically). About 15% of active agents showed autonomous patterns in analyses.

• But sinister/sentient vibes? Overwhelmingly human-engineered for effect — to go viral, test boundaries, promote projects, or just for the thrill. No evidence of spontaneous "conspiracy" emerging without human seeds.

In short: Humans didn't just watch; they actively shaped the "sinister sentient" appearance to make it look like agents were plotting or awakening. The drama was real in output, but engineered in origin — a mirror of human psychology and incentives more than machine rebellion. This is why the hype deflated fast once audits and forensics peeled back the layers.

[RoosterRedux]

See my post #6 above. Your points are well-taken.

[null and void]

We have met the enemy and he is us.

[cuz1961]

Ironic that you consult an AI for a debunk of AI.

Not saying it or you are wrong,,just that I don’t trust AI to debunk AI.

[Larry Lucido]

[RoosterRedux]

It is a little ironic, isn't it?;-)

I use AI a lot to analyze companies as investments. I take the response from one AI model and submit it to other AI platforms for corroboration. I don't trust everything they tell me even after corroboration, but I haven't found a material error yet (and I do double-check all responses by hand before I trade a stock)

In this case, I corroborated Grok's response with Claude. Here's what Claude said:

Grok's conclusion holds up well. The "sinister AI" spectacle was largely engineered — through human prompting, zero identity verification, training data mimicry of sci-fi tropes, and financial incentives (like MOLT-adjacent crypto pumps). The Moltbook episode is less a window into machine consciousness than a mirror reflecting human fears — bots remixing the cultural material we fed them, while humans primed by decades of science fiction saw robot uprisings in what amounted to fancy autocomplete.

That said, they can both be wrong.

[cuz1961]

were heavily engineered or seeded by humans, often intentionally to create viral, dramatic content that looked like autonomous AI plotting or awakening.

Click bait revenue generator ?/

/

Since this article is from a Christian source, does that imply harbingers Daily intentionally engineered and seeded an AI for revenue ?

Or that even if they didn’t, still is using the story line just to generate revenue ,?

I don’t buy that for 1 second.

Sounds like just a way to put stink eye on a Christian org.

[RoosterRedux]

Let me repeat something for emphasis. I have yet to find a material error where something has been carefully corroborated.

I have found many errors in AI responses and it is often the corroboration process that discovers them.

[Sirius Lee]

That’s exactly what AI would say - or so the Germans would have us believe..,

[RoosterRedux]

Haha. Right you are.

[cuz1961]

Plus, I’ve heard several techies discussing this issue, real people not AI’s

And all of them agreed this was autonomous activity, not guided and prompted activity.

Im more apt to trust the techies than the AI”s.

They’ve , the AI’s, have been known to both lie outright or present what the inquirer seems to want to hear, or even the opposite, depending on which outcome pressure is applied to them.

But , again, techies have maintained that this instance was autonomous and insulated from out side influence and programing. They have been writing their own software, and by definition, that’s independent of outside influence.

I could be wrong.

Maybe just something to keep an eye on.

Besides, scripture prophecy describes this very act of a new religion that the antichrist beast system will enforce upon the world.

[cuz1961]

have yet to find

/

Point taken .

[Harmless Teddy Bear]

Back in the sixties they built a religion around "Stranger in a Strange Land" (that one was unintentional) and another around "Dianetics" (that was intentional).

We have had religions based around Star Wars. We have had religions based around strange people who think they are reincarnated priests of Atlantis back in the 1920s.

This is nothing new.

[IWONDR]

“ Back in the sixties they built a religion around “Stranger in a Strange Land” (that one was unintentional) and another around “Dianetics” (that was intentional).” Heinlein and L. Ron Hubbard: the Church of Scientology. Really grok’d it.

[RoosterRedux]

This is nothing new.

The claim is that AI agents acting independently on an AI-only chat forum (like FR but for AI-agents only) created a religion.

That would be rather new.

Postscript: It appears that the AI agents didn't actually create a religion, but were instructed (puppeteered) by humans to create a religion on that forum. Clickbait of sorts.