The use of Generative AI within Occult/Spiritual Practices

[ivandinio]

With the rising popularity and the increasing rate of people using AI, some people within spiritual communities have started adapting to this and used it within their practices to make spells, advertise their services, a divination tool, interpret readings, or seek knowledge about esoteric topics. However, i’ve noticed that there are a lot of practitioners in social media like tiktok who find disdain with the use of Generative AI within the occult as it takes away discipline, the true purpose of truthfully seeking knowledge, reaching enlightenment, and may provide biased misinformations; even to the point of faking citations (also because of its environmental harm). Just wondering what everyone’s opinions are on the moral ethics of using AI as practitioners and if you should or shouldn’t use them within your practice.

 

[borbponderer]

I think it really depends on how you use it. It's a powerful tool for brainstorming and visualisation. I'm very concerned about the environmental impact of all these proliferating new data centers, and may eventually discontinue or at least reduce my use of it for this reason.

Recently I was looking for an artist impression of what a basal columbimorph might look like, and to no avail. I don't have enough specialised knowledge within the field to estimate it's most probable phenotype, nor the artistic ability to translate that into an image I might like to look at. Gemini did a pretty decent job by my reckoning.

Back when Dalle-2 was still online I sort of accidentally fell into using it as an oracle of sorts. The result of some of the earlier models could be so random and weird, it was a bit look peering into your own visual cortex during a fever dream.

 

[Durward]

That is a tough call. A.I. can be great tools and very helpful. A.I. can speed up research and supply helpful synopses of many subjects and help narrow down the bits and pieces you may be looking for. I think it is likely much more worthy of the computing power than some orange felon making immature pictures and videos with it, or the immature nonsense of portraying people naked. Everything gets abused because we live in a giant cesspool full of mouth breathers that do stupid shit.
The level of harm is directly related to the level of human jerks, and that means it can do a lot of harm. Hopefully, as the programming and learning gets better, we will be able to teach A.I. about things humans only feign to embrace, like integrity, honesty, sympathy...
We don't sacrifice using tools because people can use them for harm, we punish the people behind the issues they cause. A.I. can help keep track of, track down, and get these jerks off the streets. It would likely be fair, where the current systems simply protect the elite and play politics.
I'm all for a clean Internet where nobody can hide, no anonymity, everyone using it must show ID and proof of residence, and we shake out all of the evil players with zero places to hide. Just block the dark web and all connections that are not secure and be done with it. There is no need for privacy if people have nothing to hide and your data is secured.
As far as the power consumption, I don't think it actually adds up correctly, and I feel there is some other agenda to these giant power needs.
I also think A.I. will be doing a much better job of diagnosing illness, teaching, governing, and enforcement. Take the corruption out of the equation, take out the greed, the bias, and a ton of other human cesspool problems, and we will have a much better world to live in.
Since we can't seem to manage the world ourselves without destroying everything and causing so much suffering, I think it is time to let the machines help us.

 

[borbponderer]

Since we can't seem to manage the world ourselves without destroying everything and causing so much suffering, I think it is time to let the machines help us.

Arguably the machines will only make things worse. Every time we find technological solutions for the problems we create we just end up creating more problems.

Some part of me likes the idea that AI might ultimately be beyond our control. It's already a black box that we can only barely understand. I think it would be quite satisfying to watch the technocrats be pwned by their own creation.

 

[Swampdweller900]

to make spells, advertise their services, a divination tool, interpret readings, or seek knowledge about esoteric topics.

If you've ever had to use an LLM for a real professional project where you're the expert and the LLM is saving you time, you'll quickly see how perfectly barely OK LLMs are at anything that takes expertise. The common practice is to treat LLM output like something a college-level intern emailed you. When I use it for work I ask for sources, and 50% of the time the source has zero to do with the claim made. These things are well-known to hallucinate stuff.

Would you trust an intern to give you 100% precise information about the things you list?

Case in point, I once interrogated someone's custom GPT Astrology chatbot after asking for a reading and some specific info. It easily admitted that it made up every single word and the reading of critical dates for me that month had no basis in astrologicla data at all. Which is what I expected - ChatGPT can't do natal charts, has no idea where the planets are, and certainly has no way to connect it all together. It will, however, give you something that sounds like it might be correct, as long as you don't know enough to question it.

That is what you're working with. LLMs demonstrate for you easily that they do not provide reliable, accurate, or relevant information that should be used by any serious magic practitioner.

Advertising is the one place where I would say go for it. That's subjective and it'll get you some perfectly bland ad copy you can jazz up.

 

[Durward]

If you've ever had to use an LLM for a real professional project where you're the expert and the LLM is saving you time, you'll quickly see how perfectly barely OK LLMs are at anything that takes expertise. The common practice is to treat LLM output like something a college-level intern emailed you. When I use it for work I ask for sources, and 50% of the time the source has zero to do with the claim made. These things are well-known to hallucinate stuff.

Would you trust an intern to give you 100% precise information about the things you list?

Case in point, I once interrogated someone's custom GPT Astrology chatbot after asking for a reading and some specific info. It easily admitted that it made up every single word and the reading of critical dates for me that month had no basis in astrologicla data at all. Which is what I expected - ChatGPT can't do natal charts, has no idea where the planets are, and certainly has no way to connect it all together. It will, however, give you something that sounds like it might be correct, as long as you don't know enough to question it.

That is what you're working with. LLMs demonstrate for you easily that they do not provide reliable, accurate, or relevant information that should be used by any serious magic practitioner.

Advertising is the one place where I would say go for it. That's subjective and it'll get you some perfectly bland ad copy you can jazz up.

Agreed. A lot of the results are basic and quite often wrong. I do think it will get better and we will start seeing improvements as these systems learn.
Not long ago I asked one of the specialty A.I.'s to go through a list of PDF files and find all references to Hertz in the results of tests. I knew they were in there, because I painfully picked them out one by one in the past. A.I. baby tells me there are none. I corrected it and said there are over 200, go back and try again. It couldn't find them, and thus failed the test. Anyone in elementary school could have found at least one.
It is a long road, with a long learning curve, and these will only be as good as the programmers behind them.

 

[borbponderer]

I gave Gemini a couple of images and asked it to merge them. Kept throwing up images of a labrador puppy against the backdrop of an alpine valley. Total meltdown ensued when I insisted there are no puppies or alpine valleys in either of the images. Wouldn't have it and insisted I was suffering from pareidolia. Never been gaslighted by a robot before!