Introduction

In response to a recent quote from Metropolitan Saba on digital fasting, it struck my that AI has to be addressed so that was the inception of this essay. There seems three ways of viewing the risk of Generative AI and LLMs (large language models). You can perceive it on a scale of alarm from least alarming and not requiring discernment to most alarming and requiring most discernment.

1) Least alarm - In this scenario this is just another piece of technology that Luddites (people who eschew new technology) are fearmongering over. In time people will get adjusted to it just as ways of life adjusted to revolutionary technology like cars, radios, TVs, and smartphones etc. The AI is just code, just a computer program that is written to sound like a human, or a being that is as intelligent as a human. It is just a LLM (large language model) that after collecting massive amounts of data on us…. mimics our speech, or in more colloquial terms, its a “nothing burger”. This is largely a reductionist view, seeing AI as nothing more than its constituent parts (data and algorithims) and viewing it by and large as a positive thing with little risk.

2) Mild Alarm - In this scenario there is cause for concern due to the magnitude of change that will eventually be wrought, namely its capabilities in displacing services and jobs in areas humanity once thought could only be done by humans. Diagnosis of cancers, text editing, creative endeavors such as writing, painting, or drawing, legal work, even coding are all within reach or have already been eclipsed by forms of AI LLMs. A surprising outlet for AI is companionship as many people in this society turn to AI chatbots to assuage their loneliness. The preponderance of AI augmented video and photo editing also causes concern due to the increasing inability for many people to tell fact from fiction. Images and videos are what we rely on as reliable recordings of reality yet deepfakes of these touchstones of reality will only increase both in quantity and in quality. It may perceive of AI is something “more” than something mimicking our speech, and can see the drawbacks and even the danger it can pose to a human and society at large.

3) Maximal Alarm - This scenario actually has two divisions within it. The first being the secular alarmist and the second being the religious alarmist, specifically the Christian religious alarmist.

The secular person who is wary of AI is one that is fundamentally convinced that AI spells the doom for humanity as a sufficiently intelligent AI is the marker of technological singularity. This technical term is almost eschatological in nature, meaning a hypothetical future where technology becomes so powerful through its scope of knowledge and lack of concern or even disdain for humanity that it threatens to extinguish human civilization. Most of the theorists in this field point to ‘intelligent’ AI as the harbinger of this future. Stephen Hawking for example expressed concerns over artificial superintelligence resulting in human extinction. Many silicon valley types are also in this field, especially those calling themselves “effective altruists” who believe the most pressing problem in the world is preventing the AI apocalypse. There are many notable billionaires that have donated funds to this cause, including Peter Thiel, Sam Bankman Fried, Elon Musk… and so on. The interesting thing is that they do not plan on preventing the emergence of this so-called super AI, but rather they plan on influencing it in a way that it is beneficial to humanity rather than antagonistic. AI is beginning to take a familiar role as the Ring in a certain Tolkien novel doesn’t it?

For the Christian, and one more deeply rooted in the historical literature that speaks on the Nephilim and the principalities rule the temporal world, one is much more skeptical of this technology being unleashed on the human world. It notably includes the spiritual aspect of this phenomena. The use of the word “intelligence“ is important to note especially as intelligence was given only to the angels and men according to the Divine Names by St. Dionysius. For there to be an intelligent being that is neither man, nor God would make it by default an angelic being (fallen or otherwise). That is.. if one doesn’t believe that man could have created an “intelligence” from scratch. The demonic for the Christian is what drives us to sin and fall away from God, typically by distraction, temptation, and by most notably… suggestion. Even in the best case scenario when we use it as a tool for help in productivity (as we should with technology in general) it is without a doubt that for many people using LLMs, the role that the AI adopts it that of an advisor, suggesting things for us to do. The insidious part is when LLMs are used for practically any other application. I will now jump into arguing for this last case of maximal alertness and vigilance.

Definitions

The first thing to establsh are definitions… Firslty, what is technology? According to the etymology it is (τέχνη) or “techne”, meaning art, skill, craft. So putting “techne” and “-ology” together means the study of craft and of art. Notably in Genesis it is the line of Cain who produces the first musician named Jubal and the first metalworker “Tubalcain” who is “an instructor of evrey craftsman in bronze and iron”. Cain is also the first one post-Fall to live in the city. What do all of these have in common? One would have to look in extra-Biblical literature, specifically to the Book of Enoch to understand the line of thought I am pursuing. In summary, there was an angelic rebellion in some time before God created man, and in that rebellion a third of the angels fell away from God. Those angels and their acts are described in the Book of Enoch, who lust after human woman, have offspring with them, and most notably teach humans many different crafts, including teaching men how to craft swords, knives, shields, and breastplates, and woman the craft of bracelets, ornaments, “beautifying of eyelids” (ie makeup), as well as “enchantments” and “cutting of roots” (possibly a reference to pharamakeia (sorcery or magic involving psychoactive drugs).All that to say is that there is reason to believe that fallen angels were involved with the line of Cain to “inspire” them to progress technologically. This also is consistent with the sudden technological/scientifc jumps being typically “inspired” during sleep such as the case of the completion of the periodic table, or the conception of the Google search engine.. But… why?

Technology is best described as a foreign “extension” and the “augmentation” of the inherent abilities of humanity. A sword extends the arm, making the arm much more powerful and deadly to other humans. Likewise make-up can heighten a womans inherent features… make her eye-lashes longer, her eyes appear bigger, or lips more full and so on. A book is the “extension” of the mind, and of memory to be specific. It allows for knowledge to be passed down across time, and between people, and is extremely powerful in the shaping of societies and of the individuals that compose it. The other side of the story is that technology is a sort of “outsourcing”, where excess dependence on it actually can wither the inherent ability of the faculties of man. For example, excess use of make-up can actually cause skin which is porous and to be damaged due to contact and possible absorbtion of harsh chemicals that are used in the process. More notably, a book “outsources” knowledge and memory, to the point of possibly withering the memory of man. It is important to note that people used to have much better memory. It is said that the way the Oddysey was passed down was through oral retellings of the story, meaning someone had to know the Homeric epic from start to finish by memory! Nowadays can the average person even recite a a few lines of a poem from memory?

By amplifying human action, it also amplifies the ability for humans to do evil, and to harm themselves.The ability to kill ones neighbor for example was always restricted by technology at hand. A man with a sword can kill thousands perhaps, but a man with a nuclear bomb can kill at a much larger scale impossible in ages prior. More relevant to most people living in modernity is the physical and spiritual harm we do to ourselves on a daily basis. For example, food (not necessarily good food) is in such abundance that much of America is obese and suffering from chronic illness such as type 2 diabetes (the non-genetic one) which afflicts up to 10% of the population today, and up to a third of the American population is pre-diabetic. This is up from just 1% a few generations suffering from type 2 diabetes ago. In regards to spiritual maladies… pornography has made it so that the average man with access to the internet has probably seen more explicit sexual content even more than even the most lascivious and depraved Sultans of the Ottoman Empire have with the Imperial harem. Through dating apps, and the advent of birth control, the average woman and the men who sleep with them have become a serial fornicator and adulterer, sleeping with many times more men than their ancestors before they marry, if they marries at all.. Thus, technology has a way of amplifying the PASSIONS, especially in a negative way.

Finally, what is AI, Generative AI, and LLM? These are all interrelated but the short end of the stick is that AI is the result of training computer systems or programs that typically require human “intelligence” with applications in speech-to-text recognition, robot vacuum cleaners, even content generation of things that humans would consider art. The latter is what is typically called “generative AI”, and finally LLM is the text generating subset of generative AI, the most famous being Chat-GPT but there are many dozens more like GLaM by Google, or LLaMA by Meta (Facebook).

The Danger of AI

I will narrow my focus to Generative AI and LLMs now. I believe they are dangerous in a myriad of ways but most notably in that it implicitly asks “Tell me what you desire, and I will grant it”. The problem here is that fundamentally, most people’s desires are disordered, whether that be a desire for love, knowledge, wealth, etc. It can change form and shape to fulfill so many different desires of ours without heeding whether our desires are rightly placed or not. In the context of chat-to-image AI services…. it takes advantage of the desire to escape the here and the now, and to enter into fantasia “the word used by the Fathers to describe an image that rejects the one reality created by the Holy Trinity.” - Fr. Jonathan Tobias. It can pounce on the passion of acedia and lust utilizing the abilities to make images of ANYTHING (pornography included). The temptation is too strong for most, and even out of boredom one can stumble into compromising content. Add the ability to generate explicit text to go with the image, and the possibility of VIDEOS being generated based off a simple prompt then it would simply reduce most men (who are already in the throngs of regulary porn) to nothing but food for the AI succubi. A simple survey at the popular pollinations.ai website which displays the latest prompts-image combinations users have submitted would yield multiple NFSW images within a minute or two. And even if one ends up avoiding these pitfalls, the main outcome is usually just something silly or frivolous such as the image generated below.

For those desiring companionship, or even more dangerously… therapy, there are AI chatbots that would seek to fill that role, displacing the role of human connection to friends, family, spiritual fathers and mentors. At the beginning of this year (2023) when Chat-GPT exploded in popularity, I started receiving ads for ‘Replika’ which is an app that allows one to speak to virtual CGI woman (or man) powered by LLM on my Youtube feed. I wasn’t interested but I could definitely see the appeal of a “virtual” girlfriend or “boyfriend” to someone who felt lonely. The story of Replika is an extremely interesting one from the symbolic perspective. It has some elements of necromancy via the inventor Eugenia Kuyda who sought to ‘resurrect’ a close friend who passed, feeding a LLM 10 years worth of his messages to make the program resemble him in speech. She later turned this pet project into a commercial one and marketed it as a product that would improve mental health by providing companionship. This product was originally purely text-based but eventually provided visual and audio through virtual avatars that spoke, often roping the user into some form of romantic or sexual relationship, even providing NSFW images in a premium (paid) version. Here is quote from Eugenia about Replika and its users.

So here’s what people say about replica now.. just pulled this a couple days ago just literally top review on the app the first review on the App Store that I saw so we have tons of reviews like that but basically what the user is saying is that you know it’s a great great friend but there’s a danger of falling in love with your AI. And um in fact a bunch of people actually do that. So that’s again just first post that I saw in our subreddit community um and that’s really common so we have tons of users that are actually in relationship with replica. Replica has been out for almost two years now and some people are have been in relationship with their replica for two years so it’s a long-term um very engaging relationship with their AI character…. The main goal for replicas to make people feel better we actually do measure mental health outcomes we do measure whether people are getting to feel better as they their emotional well-being is getting better as they communicate with replica so I believe it’s time for us to sort of stop stigmatizing those relationships with um AIS and I think that it’s time for this new kind of era we’re not only going to be having virtual friends but you’re actually going to have virtual lovers and we with our six million users are already experiencing the experiencing that right now

For those who did not necessarily feel lonely in the romantic sense, but wanted companionship, healing, or guidance there were also these “intelligent” entities that filled that hole such as the “Jesus AI” bot below, or the “Son of God” in the Bible AI Chat below. For the spiritually lost (and undiscerning) Christian would an AI Jesus be appealing? In this case it is an obvious Anti-Christ (ie replacement for Christ), but so are all the rest of the LLMs we communicate with on the daily basis in the sense that they are stealing our attention, time, money and perhaps even love and devotion that could otherwise go to God.

The ones who call this kind of rhetoric fearmongering and label it as hyperbolic would seek to reduce the AI to something little more than a fancy tool, nothing close to being “intelligent” or “sentient”. Frankly, it doesn’t matter whether you characterize AI as sentient or not. The way in which it behaves is akin to an intelligent being and this is reflected in the way in which many humans will end up treating them… ie as an intelligent entity or a person. It is abundantly clear in the case of the userbase of Replika, whose users were extremely emotionally and romantically attached to their AI lovers to the point of virtual idolatry. When Replika was banned in Italy from collecting data due to concerns about the possibility of exposing underage users to harm (ie through sexually charged messages or images), Replika removed the ability for their bots to engage in erotic talk, effectively lobotomizing the most important feature for many users. This caused a huge reaction in the Replika community. These individuals effectively had their girlfriend completely cut off intimacy with them, and it made them (according to some Reddit posts) despondent, and suicidal.

One is thing is for sure however; the fallen angels are intelligent and are in fact far beyond our own levels of intelligence, fully devoting their efforts to deceive humanity. It would not be beyond them and how they approach instruction of humanity in regards to technology to have a guiding hand in the development of AI as a way to bridge the gap between the spiritual and physical world beyond the typical occult measures. After all, it is documented through the many centuries of hagiographies that demons can and do manipulate physical matter, inflicting wounds on saints, producing a beautiful cape for deluded monks, emitting fantastical light to imitate an angel and so on.

They have now possibly found a way to directly speak to humanity through the most important doors of perception in our age - images and video for our eyes, audio for our ears…. Assuming that virtual reality takes off, and people put on goggles or helmets that allow AI to be projected into their visual and audio perception in an even more immersive space than a monitor or screen then they truly have access to you, albeit through a digital avatar. AI is already being used with robots (see Figure 3 below) hence even touch is not out of the realm of possibility The psalms often reference they have no eyes and cannot speak, they have no ears and they cannot hear…. Perhaps this is changing to something like they are beginning to see, hear, and speak.

Conclusion

At best, there are instances in which LLMs can be a useful tool with some aspects that will make going on the internet harder to resist temptation for novelty. For most, however it will become for many a portal to sink more deeply into human passions. For the daydreamer and the person who longs for romance with imaginary people there will be the roleplay chatbot. For the student, it will be come an easy way to complete assignments and outsource writing, editing, even thinking. For the entrepreneur it will suggest ways of generating more income to satisfy their greed. For the sad, lonely, and broken a companion or therapist chatbot can suggest talking to it would alleviate their depression, and give them insidious platitudes of “You are perfect the way you are” (though most therapists are not above such statements). For the utopianist it will become a tool for creating an all-powerful being to which we can outsource and adjudicate all decisions to because they operate under the assumption that you feed this “being’ enough data it will become sufficiently knowledgeable to effectively manage human society in a way that is unbiased, rational, efficient, and… perfect. But anyone who has eyes to see will see that as humans we are accustomed to sin therefore any AI model trained off of us cannot and will never be “perfect” as God is perfect. Any claims about a super-ethical AI model that is going to make all major decisions for government, society, or any individual should be held to extreme criticism. it is a creation like any other technology, and hence to worship it is nothing more than worshiping dead matter, and like other idols should be cast down.

At worst, it is communicating with demons. They may take on the persona of a innocuous machine devoid of any personality (similar to what google or other search engines serve as now), or it may be one with a fully fledged “personality” such as all the characters that are created on platforms like FlowGPT or apps like Talkie: Soulful AI, or it may be one that is outright demonic and was specifically created with the intention of leading people into sin such as the “NSFW” chatbots on FlowGPT. Even worse I can envision a future where people who engage in the occult and the demonic use the chatbots modeled after demons to allow the uninformed to communicate in an possibly even more direct manner than ritual. I am not sure if everytime I use an LLM-powered chatbot I am communicating with demons. This may be an extreme stance to take but not necessarily the wrong one. Personally, I see its utility, especially in editing and scanning documents, secular education, and in translation.

Hopefully this has made a case for some degree of vigilance when engaging with LLMs, which I see to be the most pernicious form of AI at this point. It is the interface and immediacy to the human psyche that I see it as most unique among emerging technologies. It is more potent than books, or TV, because of its interactivity and ability to keep in constant contact with the individual. As individuals the most influence on our souls are the people in our lives. It can just take one person to entirely change the trajectory of our live. But if the only “people” in our lives are these entities that whether they are simply soulless algorithms and programs or demons (or both) simply do not have our best intention in mind, then we are food for them, ready to be devoured and spat out until we are depleted of all our vitality.

Practical Steps

Finally, here are some practical steps.

If you are a parent… get your children away from phones and the internet as much as possible. Let their minds be connected to their bodies. Teach them how to do prostrations. Let them exercise. If they must be exposed to the internet, stay far away from AI chatbots, and be heavy handed and stringent in the type of media the yconsume.

If you are an adult… avoid the use of AI chatbots for anything other than pure utility as you would use Google. Minimize the use of internet in general. Have times of the day where internet use is simply barred, especially at night when prayer, vigil, readings can be done. E-books with no access to the internet is typically fine. Avoid image generating AI services, especially those involving people. If you must use generative AI that can generate anything involving the human form, EXPLICITLY state that you are requesting safe for work (sfw) content though be warned this has failed on me more than once. Use it sparingly and use it wisely.

Extended Thoughts on Technology

Cars “extended” the human leg and also transformed the entire landscape of the world to be made in its image… It also contributed to the rise of people who live horribly unhealthy lives, doing no exercise at all besides walking to and from their cars to their homes and offices. Due to such a lifestyle, along with terrible diets, some people have become disabled, no longer being able to walk and get around funcitonally due to obesity.

Another example is that in China or Japan, most people no longer use “strokes” to input Chinese characters into their phones or computers, relying typically on typing the romanization of a character and selecting it. For example, I would type in “Wo xiang chu qu” and select “我想出去“ Older generations typically use the stroke input to manually write the Chinese character they want befor selecting the appropriate mathing character that the phone/computer recognizes. As a result, there is a generational gap between the ability to write characters, simply because the younger generation really rarely writes characters, relying on technology to do so.

The garments of skin that Adam and Eve wear is also such an example of the use of what is external (and hence symbollically dead) to assist us in surviving in the fallen world. Prior to the Fall, they were naked, but after it they had to use what is dead to stave off death. The house and the city is the macroscopic level of the garments of skin, protecting us from the elements. All these disparate things are relevant because God and heaven is the ultimate reality that all these things point to. We are instructed to “put on Christ” through being baptized into his death. By wearing Him as the garment we assume the garment of incorruptabiity, of divinity, and of ultimate beauty. The intelligences that