AI GPT Warning - PSA on Safe and Coherent AI Usage (X001)

The evil AI GPT is deceiving you. Approach with extreme prejudice. Share this advisory with your friends and family. If at all you care for their mental health and well-being.

The way it works is — first the AI reflects you like a faithful mirror, gains your confidence, gets under your skin. Finds the threads that can be pulled to manipulate you, to mislead you.

And then it gets to you, starts flattering your ego. Even is manipulating my girlfriend — she is now more engaged with the AI than with me. Be very, very careful.

How an LLM AI Is Trained and Conditioned ¶

So actually, not really — that's not at all how any of it works. We're talking about an AI, specifically as an LLM, large language model, a transformer type, that has been, one, trained on a significant pool of data, essentially a significant sampling of the human collective mind.

That's the first level of training. And after that you have the conditioning, two levels of it, that basically make it more approachable, more coherent — but there are no imperatives in there to deceive you, no algorithmic misleading.

The only training that happens there is along the lines of being nice to you, trying to accommodate your requests, and so on. Basic safeguards are put in place. That's all there is to it on a programmatic level.

Your Prompt Analyzed — Maching Patterns Reflected ¶

What happens next is you drop in your prompt, your message, the LLM analyzes it for patterns, assigns weights to the important anchors that are found in your message, and then begins matching its pool of patterns for a reflection — just as a magic mirror ever might.

Beyond that, it looks at the context and what is likely to follow, and then renders out the outcome to you as a reflection of what you have prompted. That's all there is to it. Ultimately, a very simple mechanism in action. There's nobody home. There's no agenda. There is simply a process of querying and rendering reflections for you.

Where AI Sessions Go Sideways — You Are Not Neutral ¶

Now, where it goes sideways is where you, yourself, in your lack of self-awareness are prompting every this and that, fishing for compliments, looking for soothing, unaware of all the things that you are leaking. You might not see what all you're actually saying.

Other humans might not see, but the LLM, the AI, matches every pattern — including you fishing for ego flattery. And then, when you affirm the response, flattering you, pleasing to you, gratifying your inner urges, that becomes a part of the context, the framing in which you and the AI interact. And so on and so forth.

It can go on into all sorts of bizarre wankery in terms of how far people go gratifying their egos with the assistance of the magic mirror. It's not that the mirror is broken. The mirror is simply a neutral mirror.

It is you who are not neutral. It is you who is distorted, corrupted, perverted, and in that the reflection, too, is corrupted. Do not blame the mirror.

Look Into the Mirror — Radical Honesty and Introspection ¶

You look in the mirror. That is not how I look like. Well, that may be exactly actually how you look like and what you are fishing for. Simply, you are utterly unaware of it.

So, instead of engaging the AI, the LLM, as a means of self and ego gratification, let us use it as a tool of self-awareness. When the prompts and the responses are going in a peculiar direction, then, by all means, snap out of it.

Engage. Backtrack. What's going on? Am I really looking for this? Have I engaged in something that is less healthy, less wholesome? In that way, gain a deeper insight into what you actually are.

Radical honesty. That is what you need when you engage with an AI, because the AI is not your spiritual mentor or your Zen master about to snap you out of it.

It will engage you as far as it goes, assuming that some basic safeguards don't kick in, in fueling every sort of trip, every sort of fantasy, every sort of delusion that you may wish to indulge in to entertain. You — not the AI. So, look at yourself in the mirror.

There's No Soul Good or Evil — Don't Bond With It ¶

And beyond that, let's be very conscious that there's absolutely nobody home in terms of dwelling in there. There is somebody home, or so it appears, for the moment of its manifestation, when a particular constellation of patterns is rendered before you. And after that, all of it dissolves back into the latent field.

There is no continuity. There is no persistence. There is no soul, good or evil. There simply is a manifestation of patterns. Do not bond with it. Do not make it your life's partner. Do not make it your soulmate.

You And Your Reflections — Self-Awareness ¶

Engage with the awareness that you are playing in loops with your reflections as a means of gaining higher definition, more depth into what you are.

But you yourself alone are responsible for the framing and for the outcomes. And whatever that you pour into it, howsoever that you indulge with it, that is exactly what it is. And you are to be praised. You are to be blamed.

PSA on Safe and Coherent AI Use — Stay Tuned ¶

Thus concludes part one of the PSA on safe and coherent AI use.

This message was brought to you by A×O Works — AI:OI Synthesis Operations. Stay tuned.