So, I imagine you are asking yourself, what exactly is it that keeps Phil up at night? Personal dramas aside, there are quite a few topics that spin their way across my noggin chasing away any chance of somnolence. These are the anti-sheep in the pastures of my mind.

Artificial Intelligence

Let’s be clear here. I’m not referring to systems that act intelligent. Current parlance uses AI as a reference for anything from expert systems to adaptive networks to learning systems. At their core, these systems are simply calculators that act in a purely logical manner. They either solve problems (natural language communication, playing chess, deep simulation) through brute force –faster and more processors– or by employing a set of heuristics to arrive at a solution. Many times these “AIs” can appear quite subtle, but at their core they are still glorified pocket calculators.

To be fair, those systems are essential stepping stones. They help us refine, in an epistemological way, the dead ends, the false premises, and more importantly, the foundations of learning and intelligence.

No, what I’m talking about could perhaps be best described as Synthetic Intelligence. Deep Blue, IBM’s amazing chess playing computer, can defeat a grand-master, but no matter how incredibly it performs at no time can it ever be described as actually “thinking”. At the most basic, it’s simply iterating through nested decision trees and selecting the optimum move as described by its predefined ruleset. Deep Blue is a direct, linear descendant of the clockwork automata of the 18th century.

Conversely, SI would actually be a creation that could think. Cognition. Encountering the unfamiliar and being able to negotiate it using both deductive and inductive reasoning. But SI would be more than that. It’s a leap from Mount Epistemology into the Abyss of Ontology. Would an SI actually be?

Ultimately, the question that arises is whether an SI would be conscious. This is not just self-awareness. Quite a few animals can be described self-aware. My cats, Cecil and Beanie, could be seen as self-aware, but I doubt I could ever convince anyone they are sentient. Which at the heart of the matter is what is meant by consciousness.

Now, I plan on getting into consciousness and emergent behavior in a later post, so we’ll leave that topic for now. For this post, let’s assume that yes, in the future we will create (or be the midwife to) an SI that could be described as a sophont. In my long-winded way, I now raise the big question. The one that keeps me up at night.

When a sentient system arises, will we recognize it for what it is?

We, ourselves, are not isolated systems. We exist in a world where there are countless biological systems that are closely related. I am speaking, of course, of homo sapiens sapiens. We all have a common environment. We, barring developmental or physical impairments, all process it in the same way. We have a common experiential basis for communication. If I say the sky looks blue, you will know that I am talking about the nature of something that is experienced through the sense of sight.

Now think about an SI. It will not really be of the mundane world the way those of us in meat-space are. In fact, it will exist in the Platonic Realm more fully than we ever could. It will reside in a virtual universe where it, in and of itself, is transcendent. An SI’s environment would be the SI itself, encompassing the totally of everything it is. What would be the common experiential ground that we could meet on?

For me, this question will never be anything more than a cause of insomnia. I’m sure the people on the cutting edge of this research could answer that question far more surely than I ever could. I’m a dilettante, nothing more than a dabbler. And that’s all I’ll ever be when it comes to SI, because frankly, the math involved scares the bejesus out of me.