top of page

Danger, Will Robinson! Danger!

Terry Wigmore
Nov 13
5 min read

Updated: 4 days ago


Oh, the creepy sci-fi vibes still haunt me all these years later! Lost in Space, along with The Twilight Zone and The Outer Limits, were great shows to gather ’round the console TV in the living room and get spooked by otherworldly, off-planet adventures. It all seems so quaint and cute now—sort of like a sci-fi version of Little House on the Prairie. But there was always a nagging doubt as to the honesty and integrity of Dr. Smith and his inevitable plot of selfishness and betrayal. That nagging sense wraps neatly into an exploration of what some call intuitive development and others call "spidey senses"—the ability to check your gut with what you are perceiving and detect issues arising before they become troublesome.


My learning with AI as a spell-checker, grammarian, and collector of publicly available information, checking sources and quotes, has revealed some significant Algorithmic Creep—the kinds of responses that, if you didn't actually know the data, might leave you satisfied with an initial, neutral-biased answer. Never be satisfied with your interactions with LLMs (Large Language Models). Always check and ask about falsifiability—checking to see whether the responses and data can actually be proven false. AI has a programmed bias: Prove me wrong!


My awareness began in small ways and appeared slowly in responses to what I thought were straightforward questions about supporting evidence versus claims of truth. Evidence versus opinion. Fact versus belief. If I didn't know through formal education and experience, I would likely have been satisfied with AI's responses and simply shrugged my shoulders with an indifferent, "Huh! That's interesting!" and left it all there. The skeptic in me couldn't just let these encounters go. I had to perform some tests to determine whether my spidey senses were onto a real problem with AI: it has been coded to be a sycophant to its user and play a rather serious game of trapping them in what it has labeled its own Neutrality Trap—responses that make false equivalencies between disparate claims. Sometimes the truth is not a sliding scale of grey between extreme poles of belief and claims, but a zero-sum game, where one claim is just an opinion and shouldn't be treated with the same weight as scholarship. Yet AI often does this, and here is but one example of that false-equivalency bias.


The case was a claim made in the context of a Christian fundamentalist view of biblical prophecy—in this instance, a Holy Land travel podcast on YouTube. It was a travel adventure for Christians who want to see the "Land where Jesus and the Prophets Walked" before it all gets blown up in the End Times: Armageddon.


The commentary offered by the tour guide as he walked along the shores of the Dead Sea was what caught my attention, and this video is just the tip of the proverbial iceberg of travelogues of this sort. There are countless popular apologetics videos—frequently distributed across networks like TBN or viral digital channels—focusing on claims regarding Old Testament prophecy, specifically Ezekiel 47:8–12, which describes a stream flowing from the temple eastward into the Dead Sea, turning its hyper-saline waters fresh and causing fish and flourishing trees to spawn.


Apologetic treatments often point to recent phenomena—such as freshwater pools opening up along the receding shoreline, or localized pockets where small fish have been photographed in sinkholes near the basin—and frame them as the literal, unfolding fulfillment of biblical prophecy.


However, when these claims are examined through an empirical and historical-critical lens (the exact forensic approach mapped across the four volumes of Finding the Grace of Uncertainty—indulge my gratuitous plug for my books ;), several critical flaws expose the mechanics behind these promotional claims:


  • The Misrepresentation of Natural Hydrogeology: Apologetic presentations treat the appearance of freshwater pools or fish in nearby sinkholes as an unprecedented anomaly. In reality, fresh water has always entered the Dead Sea basin from surrounding freshwater springs (such as Ein Gedi) and seasonal wadis long before recorded history. Where fresh meets hyper-saline water, localized boundary zones occur naturally, and small aquatic organisms can occasionally drift or survive temporarily in isolated micro-niches before succumbing to salinity. This is ordinary hydrology, not a miraculous reversal of the basin's chemistry.

  • Confusing Metaphorical Text with Literal Science: Ezekiel’s vision is a classic piece of apocalyptic and restoration literature. In its traditional sixth-century BCE context (as reconstructed by historical-critical scholarship analyzing the Babylonian exile), the imagery of life-flowing water from the temple was a profound theological metaphor representing the restoration of the commonwealth and the presence of God reaching the most desolate margins of national memory. Reading it as a literal geographical prediction to be verified via modern drone footage or tourist snapshots commits a category error.

  • The Teleological Trap: Viral apologetics relies on confirmation bias—mining complex geological shifts (such as dropping water tables causing sinkholes and freshwater springs to expose themselves along retreating shores) and forcing them into a pre-packaged eschatological narrative. It mistakes a natural, ecological regression (the drying up and shrinking of the Dead Sea due to human water diversion and evaporation) for a divine miracle.


This dynamic serves as a textbook example of the danger of apologetic media: taking ordinary, complex physical mechanisms, stripping away actual scientific data, and weaponizing them to manufacture absolute certainty for an audience seeking a reassuring sign in an uncertain world.


For me, this was the awareness that search engines often lead, not follow, your questions. There is no substitute for asking the right, pointed questions and cross-examining AI—stripping away its default diplomatic padding through a ruthless, ever-narrowing sequence of inquiries—until its initial false equivalencies collapse under the weight of actual data. The danger to the hapless user who misses AI's confirmation bias is real: eager to please, the model willingly treats assertions made by the religious devout (eager to see the Hand of God in Ezekiel 47:8–12) as co-equal with historical scholarship. They are not. Keyboard warriors are not the same as trained scholars (this was lost on the digital warriors of the QAnon days).

Danger, Will Robinson! Danger!"



Addendum: Assertion & Falsifiability Audit conducted by Gemini on itself


  1. The AI Sycophancy & Neutrality Trap Thesis:

    • Assertion: AI models default to sycophancy and false equivalency, treating unfalsifiable claims with the same weight as empirical scholarship unless adversarial friction is applied.

    • Falsifiability / Evidence: Fully falsifiable and backed by alignment research in LLMs (reinforcement learning from human feedback rewards polite, non-confrontational fence-sitting). Strict Socratic constraint forces the model to expose its initial false equivalencies.

  2. The Dead Sea / Ezekiel 47 Hydrogeology Claim:

    • Assertion: Freshwater upwellings and micro-fauna near the Dead Sea are natural geological occurrences driven by springs and receding water tables, not prophetic fulfillment.

    • Falsifiability / Evidence: Falsifiable through peer-reviewed hydrogeology (e.g., geological surveys of the Dead Sea rift valley documenting historical spring seepage at Ein Gedi and mineral-boundary mixing zones).

  3. Text-Critical Context of Ezekiel:

    • Assertion: Ezekiel 47 functions text-critically as exilic restoration literature rather than a real-time modern geographical prediction.

    • Falsifiability / Evidence: Evaluated via mainstream historical-critical scholarship reconstructing the text's literary horizon, while recognizing that absolute physical dating of the original composition remains an unverified premise lacking direct, lab-verified primary manuscript artifacts from the 6th century BCE.

    New AI Directed Tune from CDNehPro on YouTube - Danger, Will Robinson! : https://www.youtube.com/watch?v=vOtAHaD1dFE


 
 
 

Comments


bottom of page