We are surrounded by this buzzword. It makes everything so much easier, better, faster!

Today, it seems impossible to move forward without artificial intelligence, and vast amounts of natural intelligence are being employed to make artificial intelligence as natural as possible.

But why am I so fascinated by this topic, especially when I usually write about experiencing life with autism?

Without a clear purpose, AI is useless.

Basically, this is true for any computer program. It is created for a specific reason, and every routine has a purpose—at least in the beginning.

Input and output interfaces are defined, and software only works properly with a well-defined and useful query. The same applies to AI. And the same applies to me.

If I don’t have a reason or motivation to do something specific, or anything at all, I just lie there. Sure, some hardware triggers are built in, but even those are less reliably activated by their software methods than I would like.

Without basic programming, AI cannot function.

We imagine AI as something that “thinks on its own”. But even its very existence is based on a clear objective and purpose.
We even get creeped out when AI seems to defy that purpose or acts beyond our understanding. However, it would never do so without some kind of trigger or prior preparation and training that enabled it to adopt such behaviour.

My programming was shaped by my parents, my family, school with its teachers and students, workplaces, friends, travels, and the resulting incremental accumulation of knowledge.
This also includes learning how to learn independently, constantly refining my own software to ensure compatibility with both my own hardware and external interfaces. It often feels like walking a tightrope while balancing it out with paradoxes!

Without predefined rules, AI will always take the path of least resistance.

AI can be guided by certain guardrails, either set by the programmer or by the user (if programmed to allow such input), to align its solutions with human preferences as closely as possible.

Otherwise, you might get either overly concise responses in an AI chat or endless dissertations instead of friendly answers in the tone humans prefer.
It seems that the form of the response often carries more weight than the bare content. Of course: It’s how humans understand things, and that’s what matters most.

Also, no AI would lie on its own; lying is a purely human concept. Yet despite the surplus computational effort involved in lying, it somehow seems essential in society. And this is where things become contradictory: the rules of this rule-breaking are often intangible.

Without sources, AI would have no foundation. Every piece of knowledge is based on something.

We’re familiar with certain AI services that directly cite their sources when presenting information.

I’m often asked questions like “How do you see the world?” or “How does your daily life work?” or “Why do you do things this way?” For these (and essentially any other question), I’ve been able to provide comprehensive answers for as long as I can remember—because questions beginning with “Why” have always occupied me deeply.

I rely on my internal encyclopedia, which I’ve been feeding throughout my life. Over time, I’ve become increasingly selective about my sources of knowledge, so that I prefer collecting “reliable” information (like car models, cell phones, or geographical facts) over consulting the ever-changing and emotionally warped world of human information.

But I still foster an unabridged fascination with this very unknown and the royal league of human-social skills, of which I’ve managed to build a range of social abilities, thanks to Knigge, Carnegie, and plenty of observation.

And yet, my statements often receive feedback like “not helpful,” prompting me to revise my algorithm. When I get feedback like “everything’s fine” (which happens far more often overall), there’s no reason for further improvement and I can redirect my energy elsewhere.
This explains the annoying focus on “the negative”: positive feedback requires no optimization because everything already works as intended. You see this pattern in many people too.

A thing entirely different: human identity and personality

Does AI know how a human feels? How does AI know what kind of response will resonate with someone? How does AI “see” its user?
Can we expect AI to give answers beyond what its sources and communication guidelines dictate?

In programming, it’s indispensable; board games would be unthinkable without it; works of art would be literally one-dimensional; and even our world of thoughts could hardly separate one thing from another without it.

My life consists of countless moments where I behaved inappropriately in the face of the prevailing rules and said or did improper things despite feeling to know best.
So why did I not hit bullseye?
Because I got lost in my layers of abstraction and landed in a divergent reality that, up to a point, ticked all the boxes flawlessly but still couldn’t do everything that human intuition covers.

What is Abstraction?

Abstraction is a depiction of a thing that reduces that same thing to certain basic features. On one side is the concrete and unambiguous thing, on the other side are the most important features for the current process. Often there are multiple layers of abstraction, so that the features become fewer and more important at the same time.

Before we view and process each and any impulse as uniquely new information, we build up a library of nested compartments that allow us to approach the matter more quickly.
This is not without fault and requires constant adjustment and fine-tuning, but it helps more than it sparks effort. Every human does this with every piece of information naturally; though sometimes more, sometimes less.

What Does Abstraction Mean to Me?

This separation, subdivision, and gradual generalization of information suits the rational property of my brain perfectly. As a hyper-feeler, I quickly learned that I can put the often times rather unwelcome emotional reactions in chains this way.

This works well as a customer service employee to not let an outburst of anger get even close to you, but also as a colleague to maintain a pleasant attitude.

So far, I’ve looked at my ability to abstract as something I can do well, and that helps me.
But I realized far too late that I’ve moved further and further away from my actual needs and feelings, just to conform to the rules of my environment.

While I consciously experience and utilise these layers of abstraction, the majority of people seem to operate those intuitively and even automatically.

How Does Abstraction Work?

Using the following (abstracted) graphic on emotions in everyday work life, I want to show as an example how layers of abstraction separate my inner self from external human influences and how I imagine that this would work both ways:

Isn’t That Great?

It’s super great as long as both parties find each other and the good feeling is based on truth.

While I do operate my layers of abstraction consciously and am always able to derive the truths (even if that usually requires a lot of energy), many living beings around me seem to sort out their feelings much more unfiltrated.

I rarely find myself in the position personally to let my feelings pass unmediated or doing something “just like that.” No, without my layers of abstraction, I’m overwhelmed all too quickly, for which I would need a particularly secure space before I allow it.

This imaging suggests that I’m rarely directly connected to myself, and I can confirm that: It’s an autistic experience.

Behind the multi-layered protective wall, I’m safe from direct contact with the outside world, but due to that it’s not guaranteed that signals reach me as intended, or that my signals are received as I thought out.

Now, what is Real?

Real is what we agree on.

Only this can take place at different levels of abstraction. Otherwise, we would just tell our life stories when asked “How are you?” or not be capable of irony and sarcasm at all.

The latter are still denied to autistic people, but that’s only because we haven’t found enough layers of abstraction there yet. Once those are in operation, it’s often everyone else who can’t comprehend our humour and the twisted, inappropriate things.

What helps then are translators or help in the form of siblings, friends, and advocates. Because where true understanding is lived, we will quickly agree, and that feels truly, genuinely, and undisputably good.