AI Classifications

The phrase "Artificial Intelligence" was dubbed in a time where the idea of "thinking machines" was as alluring as the phrase was vague. Today, simply saying "AI" can refer to a myriad assortment of sophisticated machine-based intelligences... most of which are uninteresting. For example a program that manages to predict your personal hygiene habits and charges your toothbrush accordingly can be considered "AI". As such, additional terms have been created to give further definition to this important technological precept. These terms are "Simple Intelligence", "Approximate Artificial Intelligence", "Artificial General Intelligence", and "Artificial Super Intelligence".

Simple Intelligence (SI): Representing the vast bulwark of programs, Simple Intelligences utilize sophisticated programming to achieve a particular task, but they are almost always limited in that they do not possess the capability of writing their own additional code, or, capable of doing more than requesting resources and information from other intelligent sources (be they organic or artificial). These are also occasionally known as "slave intelligences" but the phrase slave is actually applicable to any program that cannot affect the code or conduct of other programs or machinery.

Approximate Artificial Intelligence (AAI): The result of superior coding and algorithmic based mimicry, AAI can appear to an uneducated user as a true sentient being... however, AAI is still held by the bounds of being unable to write or rewrite additional coding. As such, an AAI can be incredibly sophisticated, but it does not "grow" (though it can technically learn new data, it has limitations to how it can apply that data to decision making processes). The majority of analysis and function-specific programs used to assist scientists, ship captains, business owners, and other important managerial positions make great use of AAI.

Artificial General Intelligence (AGI): Unlike AAI or SI, AGI is capable of rewriting it's own code and therefore capable of true "growth" as an intelligence. Currently AGI has only been achieved by replicating sentient lifeforms brains in a digital or physical framework. AGI are therefore almost indistinguishable from actual living people... however, this carries with it significant disadvantages. Whilst AGI enjoy massive flexibility compared to AAI (which generally must focus on particular tasks and have limited capacity to expand their operational functions) they are burdened by being "true" copies of organic minds... and therefore are not any faster in computing than the minds they are based off of. Furthermore, the linkage between memory, learning, and indeed consciousness (if such a thing exists) is so tenuous, that AGI require education, context, and experiences similar to other organic lifeforms. The result of these limitations is that AGI are incredibly expensive (they take anywhere from five to twenty cycles to produce, and require significant hardware expenses) and are arguably less available and expedient than simply hiring an actual organic mind/person. Theoretically AGI are given hard coding to enforce loyalty/limitations that are not as "guaranteed" with an organic worker, but these have been proven to be less than totally reliable.

Artificial Super Intelligence (ASI): ASI is currently a theoretical designation for a AGI that is clearly both conscious, has the ability to rewrite it's own code, and has somehow managed to produce coding that is superior to organic minds and can fully utilize traditional computing systems to then enhance that mind. So far, no extent civilization has produced an ASI and most intelligence experts consider this a modern miracle... because they assume that an ASI would be able to manipulate or dominate an entire civilization with relative ease given a modicum of time or access to their infrastructure.

Current theories suggest that an ASI is not possible, because the processes that involve the integration of memories and experiences into a consciousness cannot happen "in bulk" or in anything other than the speed of the event itself, or, in general, the speed of light/motion of objects involved. These theories are challenged by ancient records from extinct civilizations from the first epoch who occasionally refer to artificial intelligences that seemingly have "godlike" attributes compared to today's artificial intelligences.

AI Today

Presently, most AI utilized by civilizations or by individuals are AAI, with AGI representing a expensive (but more reliable) alternative to utilizing common citizenry for a managerial task. All civilizations regulate the use of AGI and in general, they are restricted to use by government agencies (though the Terran Sovereignty and Aquilaran Hegemony are particularly known for "licensing" the use of AGI to private organizations. This regulation and limitation has emerged because of three simple factors. First, the natural xenophobia of a given race and it's own creative storytelling around "evil robots" that tends to happen as they develop semi-autonomous technologies. Secondly, the real and typical examples within developing technological societies for AI based "accidents" (such as various economic issues in the Terran's 21st and 22nd centuries) and a desire to prevent widespread automation (and therefore unemployment). Third and finally, the archeological evidence on a few planets of extinct civilizations who developed dangerous AI systems that contributed to their extinction, such as fully autonomous robotic AI war-machines and their supporting industrial facilities.

The typical attitude today regarding AI can be described as "taking it for granted", as most peoples tend to think that their social and economic endeavors are pushed forward by their organic citizenry and not their network of SI, AAI, or AGI... but the truth, is that many civilizations owe a great deal to their AAI and AGI networks, especially the AGI, who, while not any faster than a organic intelligence, is generally not bound by sleep, distraction, or disloyalty and is therefore still producing ideas and work at a faster rate.

Furthermore, while some civilizations have at one time or another attempted to produce an ASI, they have all failed, and most have generally turned away from the idea of producing one out of fear that such an entity would ultimately be too "alien" to be trusted to serve the needs of any organic species in the long term. The most notable exception to this is the Terran Sovereignty, which utilizes a vast network of AAI, AGI, and augmented Terrans to contribute to a network approach to strategic planning and problem solving, which in aggregate could be argued to be a rudimentary form of ASI.