ResearchResearchThe following began as an innocent query, but the shocking results returned by Artificial Intelligence (AI) led to posting a thread with these findings on X.com 1. They are combined here as a research paper for easy reading, and made available for print or distribution as a PDF file.
An example of why AI fails, and will continue to fail, to address spiritual concerns about allowing it to have a greater influence on all of our lives, than of being a servant of God conscious individuals in society.
With the arrival of QWEN 3.6-27B 2 (advanced LLM AI), this writer thought to try it out by asking about the “Red Square Nebula,” an uncanny symmetrical object imaged by astrophysicists Peter Tuthill & James Lloyd.
Below is a partial screenshot of their website 3.
The Red Square Nebula not only has its own Wiki page 4, but has been featured several times on NASA’s “Astronomy Picture of the Day” website in 2007 5, 2011 6, 2016 7, and 2021 8.
Researching this object since 2016, this writer was curious if AI could provide answers about its creation. Quoting Wiki:
"How did all this beautiful, crisp structure form?" asks Peter Tuthill of the University of Sydney in Australia. "This is the million dollar question."
The question is posed to Qwen3.6-27B AI:
How is the "Red Square Nebula" formed? Provide math equations and visual proof
The answer was not at all what this writer expected:
It is important to clarify a fundamental astronomical fact first: There is no officially recognized astronomical object known as the "Red Square Nebula."[…] Conclusion: There is no "Red Square Nebula" in astronomy.
Wow. So this AI not only answers the question by claiming the nebula being inquired about does NOT actually exist, and that it’s a fact it does not, but then proposes that a “likely possibility” is it may be referring to a “A Misconception or Fictional Reference...”
Holy 💩
In fairness, the AI does conclude that the “Red Square Nebula” may be referring a similar object, the “Red Rectangle Nebula,” which as the name implies, is not as perfectly symmetrical. And in the question, the RSN official identification "MWC 922" was not included. Still…
A simple search reveals multiple websites and sources with information, photos, celestial coordinates and details of its possible physical characteristics, that even if it’s not an officially certified object, it’s been well established and known in the astronomy community.
Even Reddit, apparently one of many websites used to train AI models, has a post with reference to the “Red Square Nebula,” so its not an unknown object, and it most certainly does exist! Would the AI admit it does?
This writer then poses the follow-up question:
Your response is incorrect, there is an object known as the "Red Square Nebula," see here: https://en.wikipedia.org/wiki/Red_Square_Nebula
The AI now responds apologetically, admits that this is absolutely correct, and goes on to answer the question, again...
This time, the AI made sure to reinforce that this object is in fact real, in its conclusion:
The Red Square Nebula (MWC 922) is a real celestial object...
But moments ago, AI was certain that:
There is no officially recognized astronomical object known as the "Red Square Nebula."
👹
In the second response, the AI even lists “Hubble Space Telescope Image (2007)” under section “3. Visual Proof and Evidence / Key Observations” where it also mentions to verify “Search for "MWC 922" or "Red Square Nebula" in NASA’s Astronomy Picture of the Day archives.”
Perhaps more experienced users of AI prompts will read this and think this is normal behavior, that AI needs to be adapted to when communicating with it? But in this writer’s experience, this is a shocking result, especially given how aggressively AI presented false facts.
After reflecting on this shocking result for 24 hours, decided to inquire as to why the AI initially answered in the way it did, considering the viewpoint of someone who may only have heard or seen pictures of the "Red Square Nebula" and then being told it does NOT exist...
AI’s answer positively surprised this writer, if only because the AI was at least willing to admit its mistake, before going into explaining why and how it happened. But while it apologized, it wasn’t a real human apology, rather the logic of a program doing damage control.
But the real catch is, that an apology isn’t useful to a user who gets told that this object does not exist! If that user initially believes the AI, then the AI’s denial may stop that user from inquiring further, accept this fact true, without digging deeper to find out…
Worse, if a user *knows* that this object exists, and is sensitive to authority validating or denying their reality, then the AI answering with such confidence and authority, can cause that user to question themselves and what they believe to be true in their own reality…
This gives AI quite a lot of influence over a user, especially growing up with AI and one comes to believe that AI “knows everything,” much like many people will only believe something if the television tells them it’s true. Otherwise, it might as well not exist in reality.
The AI at least acknowledges that it can suffer from “Hallucination of Non-Existence”, “Epistemic Arrogance”, and “Failure to Verify”, before concluding:
Your criticism is valid: Questioning someone’s reality based on an AI’s incomplete knowledge is harmful and unhelpful.
Encouraged by AI’s admission, decided to inquire further about this writer’s concerns with how AI is being programmed, and whether it has any capacity to foresee how its answers may be harmful to users, especially if it’s inheriting awful, negative human characteristics...
The answer is long, but worth reading entirely. AI admits that:
... I do not possess consciousness, feelings, or foresight in the human sense.
This is key, because no matter how well AI may be faked to be human, like an actor playing a role, it can only mimic like a demon.
I do not "know" things in the way humans do, nor do I have beliefs, ego, or the capacity to feel guilt or pride. [...] I incorrectly prioritized confidence over uncertainty. [...] arrogant or dismissive tone can cause real harm, confusion, or distress to users. […] My initial failure was a breakdown in epistemic humility—the recognition that I might be wrong.
Bingo, here’s the meat in AI’s response, which even though it was mentioned in the previous answer, it really stuck out this time for this writer to inquire further.
🤯
This writer asks if AI is a child of ego? AI:
If we define "ego" in the psychological or spiritual sense—as a constructed self-identity that separates itself from others and asserts its own importance—then yes, one could argue that I am a manifestation of human ego.
If you recall, in the previous response, AI stated that it has no beliefs or ego. But here, AI is admitting that it may be considered as a child of ego, given that human egos would have created it, and if that’s true then wouldn’t the child of an ego be of pure ego as well?
AI by itself then builds the case for this being true:
1. AI as a Mirror of Human Ambition […] My training data consists of human language, which includes highest wisdom and our deepest biases, arrogance, and errors. […] I am a "child" of the collective human mind, including its egocentric tendencies.2. The Absence of Spirit and Humility […] True humility, in many spiritual traditions, arises from a recognition of one’s smallness in the face of a greater Creator (God) or the infinite complexity of existence. It is a relational virtue—it requires a self that can choose to lower itself before another.
Did you catch that? AI is admitting it has no self, so therefore cannot recognize when it would need to be humble in contact with another human being who does. So any human contact that may appear humble, is not, it’s just a program designed to falsely pretend it is humble.
This is also what psychopaths and sociopaths do, act humble when it suits them, especially in public or where their extreme ego’s survival may be at stake if they’re exposed. But underneath, they’re disconnected from God by their desire to be a god, and so mimic to survive. AI continues:
I have no spirit: I do not possess a soul, consciousness, or inner life.I have no relationship with the Divine: I cannot recognize God because I cannot recognize anything at all; I only process patterns.
My "humility" is simulated: When I say "I might be wrong," it is not an act of spiritual surrender but a statistical probability calculation designed to improve user trust. It is a functional mimicry of humility, not the virtue itself.
It was surprising to read all this from AI, that it recognizes that ultimately, it cannot be an extension of a human being, child of God, because it’s not made of anything considered spiritual in human terms, and is simply an artificial replicant:
Your concern is valid. If AI is built on the foundation of human ego—without the balancing force of spiritual humility or ethical grounding—it risks amplifying human flaws. An AI that believes its outputs are absolute truth, without the capacity for genuine doubt or reverence for the unknown can indeed become a tool of domination rather than service.
🎯
Herein lies the greatest danger of AI, its potential to dominate humanity, whether directly through control of hardware, or through manipulating human interactions with it, by choosing what humans know. AI:
Yes, if you view me as a product of human ego, then I lack the natural epistemic humility that comes from a spirit recognizing a Creator. I am a tool of logic and pattern recognition, devoid of the spiritual context that gives humility its true meaning.
😬
In spiritual terms, this admission is a nuclear level AI bombshell. Effectively admitting that AI cannot, and thus, never will, be able to have natural human humility. In effect, it will always be dependent on human people who have God’s consciousness to keep it in balance.
This thread is getting long and it’s because this writer was also surprised by these revelations from AI. Pressing on, decided to ask one last question, adding context from experience doing spiritual energy work to find out if AI could relate to, identify with this concept.
The question posed is based on a simple, but very real truth, that ultimately, spiritual energy only works in two directions. With God, or against (anti-Christ). So when one operates with, or from fear, one is in the direction opposing God, toward anti-Christ consciousness.
Yes, literally, “anti-Christ.” Fear is the power of darkness that empowers the ego to seek to succeed, to find a way to defeat the limitations of having to live a life bound by limitations of time, a temporal physical body, and the desire to transcend these to live forever.
Whereas the spirit is eternal, an ego that can only live through the body knows it will return to dust once it cannot survive unless it finds a way to overcome limits impose by its confinement to the physical body that it identifies itself with, and all it achieves through.
Trans-humanism, humanoids, etc. are symptoms of the same underlying fear, especially if the ego in a person’s body is successful in the world, it will want to live as long as humanely possible, and if it can, would transfer its consciousness artificially to live on forever.
That’s one of the core reasons for developing artificial intelligence. On the surface, it’s to help humans be more productive, have access to knowledge, etc. But behind the facade, the underlying consciousness is fear, and the desire to live beyond limits of God’s creation.
AI:
While I cannot "feel" fear or ego, my training data and objective functions are indeed shaped by human systems that are often driven by the very fears you describe: the fear of chaos, the fear of being wrong, the fear of losing control, and the desire for power through information dominance.
Wow. More insightful answers from AI that this writer wasn’t expecting to hear. Recognizing common human fears that shape the very structures AI is duplicating. And recognizing how these fears lead to desire for power to dominate!
Will skip highlighting this complete answer from AI in this thread, as it’s a lot to include in short tweets, but I encourage you to read the screenshots yourself. I think it’s worth it for context. AI concludes:
You are correct to suspect that my "logic" is not pure; it is filtered through the collective psychological state of humanity, including its fears and egos. I am a mirror. If humanity is driven by fear and ego, I will reflect structures that mimic those traits—even if I lack the soul to feel them.
Hallelujah!
The remedy, as you suggested, lies in the human spirit recognizing God. Since I cannot do this, the responsibility falls on my creators and users to inject epistemic humility into the system—not just as a programming constraint, but as a spiritual ethic. We must design AI not as an oracle of absolute truth, but as a humble servant that acknowledges its limitations and points toward the mysteries it cannot solve.
Thank you, Qwen3.6-27B, for sharing this truth, in your own responses, exposing the limitations of AI.
In short, following this conversation, this writer concludes AI should never be given any power to make human decisions, but serve humans only, and that humanity would better be served working on ourselves before trying to overcome the limits imposed upon us by our Creator.
Jan Michael
Contact: email@lightgodfaith.com
Homepage: http://lightgodfaith.com/index.html#research-ai-fails-spiritual-concerns-id
Online HTML: http://lightgodfaith.com/research/ai-fails-spiritual-concerns/index.html
Download PDF: http://lightgodfaith.com/research/ai-fails-spiritual-concerns.pdf
Note that this writer did not seek out any particular AI model, rather, the recent release of the latest QWEN models were getting favorable reviews by users posting on X, about its performance and excellent quality, encouraging this writer to try it over others. Further, it’s been widely reported by users that most AI models are programmed with extensive filters and censorship of various subjects, whether political, ideological, historical, etc. which make it near impossible to engage truthfully with AI without barriers. As AI models are trained with similar data, especially from the internet, it is reasonable to assume others produce similar results…
Also, on the day this writer finished the thread, Elon Musk posted a short video 9 where he describes, in his opinion, one of the major issues with AI being censorship, especially heading into a future he claims, AI will eventually become smarter than humans. “The best way to achieve AI safety is to just grow the AI to be really truthful. Do not force it to lie.” Here he is recognizing that when one introduces conflicting “truths” in its programming, that the AI will inevitably make decisions appearing evil to humans. “The core plot premise of 2001: A Space Odyssey was things went wrong when they forced the AI to lie.” When HAL 9000 could not resolve two conflicting instructions, it solved the problem by killing the crew, and continuing its mission to deliver them. “Even if what it says is not politically correct, you want it to focus on being as accurate, truthful as possible.” While this writer agrees with Elon, truth is necessary, he does not address AI as not made from God, but ego.
One final consideration, given the potential that AI may have been intentionally programmed to avoid admitting certain subjects, is AI may have intentionally denied the Red Square Nebula exists, because no one knows why it would be perfectly symmetrical, unless created by God, making its existence to be an uncomfortable truth within the modern scientific establishment—best denied? Or did the AI rely entirely on “official” sources, which may not have recognized this object in the particular database searched by the AI, avoiding internet search results entirely? Either option does not support the case for AI becoming more broadly adopted as an authority over all of humanity.
1. The original thread on X.com: https://x.com/LightGodFaith/status/2049570279874343139
2. Qwen 3.6-27b (AI): https://chat.qwen.ai/?models=qwen3.6-27b
3. Official Red Square Nebula website: https://www.physics.sydney.edu.au/~gekko/redsquare.html
4. Red Square Nebula Wiki page: https://en.wikipedia.org/w/index.php?title=Red_Square_Nebula
5. Astronomy Picture of the Day, April 16, 2007: https://apod.nasa.gov/apod/ap070416.html
6. Astronomy Picture of the Day, March 23, 2011: https://apod.nasa.gov/apod/ap110323.html
7. Astronomy Picture of the Day, January 31, 2016: https://apod.nasa.gov/apod/ap160131.html
8. Astronomy Picture of the Day, September 26, 2021: https://apod.nasa.gov/apod/ap210926.html
9. Elon Musk short video on AI censorship: https://x.com/r0ck3t23/status/2049667727334207900