Artificial Intelligence (AI) exists. This is certain, and demonstrable. DARPA sponsors contests that utilize AI to navigate a course using an unmanned vehicle without a remote control. However, this is what is referred to as "narrow AI"; it's objective oriented, and while it is capable of independent action, it is by no means capable of what we would consider "thought" or "free will." Digital Self-Awareness, i.e. a computer that has an actual sense of identity, exists in science fiction, with mixed outcomes. In some instances, it bonds with its creators; i other instances, it wipes them out. In any case, the rapid advancement of human knowledge into neuroscience, psychology and computer science will generate both a "wide AI" and self-awareness, the latter possibly by accident. If indeed self-awareness spontaneously generates, it could have a disastrous effect on the world.

An option to preempt a disastrous "awakening" is to design self-awareness in a device that is isolated from the internet. But there are many philosophies, and a key question is; What can tell us whether a computer design has achieved a true identity, not merely a programmed one that mimics our own consciousness? In 1950, Alan Turing had a test for Artificial Intelligence. It was simply this; can a real person speak to the machine and not know they are speaking to a machine?

This is a good test, but in advanced self-awareness prototypes, this test becomes very subjective. We need to establish a better set of tests. But in order to do that, we must understand what we're testing. Assuming we want to replicate human psychology in a machine, we need to know what the core properties of being a human are. 

There are three essential philosophies for understanding self-awareness in humans, with both physical and non-physical corollaries. The first philosophy is "roots up." A "roots up" philosophy depends more heavily on physical mechanisms, and deals in early fetal- and child development. It follows that by creating "birthing" conditions (or their digital counterparts,) an understanding of self-awareness can be reached. One example of this would be how a child can't be deceptive until about the age of three or four. This skill, aside from it's negative connotations, is key to the survival of the human psyche, as well as, in some cases, physical survival.

The second philosophy is a "branches-down" philosophy. It deals more with the non-physical, but has its own physical counterparts. It looks at the work of psychologists, concentrating at the matured mind, and explores that mind to develop key understandings that can broaden the general understanding of digital self-awareness. An example is the miniscule lapse between the time sensory information hits the amygdala (emotional brain) and the time it hits the pre-frontal cortex (logical brain.) Our emotions have a slight lead on our logical judgment. This is an important insight, and should be reflected into the self-aware system design.

There is a third; perhaps we could call it "the forest", that deals with the power of the situation to affect behavior, as well as group roles. A self-aware system must be able to react to its environment, and must be internally altered (re-wired) by its experiences in that environment. Future decisions would not only be affected by input, but interpreted by the filters of past experience.

Creating Full-Spectrum Artificial Intelligence with Self-Awareness will require a multi-disciplinarian approach. This is difficult in an age where detailed knowledge is so specialized, but creating a self-aware system that is under human control is a valuable hedge against one in the future that springs up unexpectedly in a military supercomputer.

Show (0) - Add comments: