AI Classification
AIs (Artificial Intelligences) pervade large parts of the world, and particularly the internet and infrastructure. Few governments don't have numerous AIs for various purposes. Most superhumans, heroes and villains, have them as well, even if they aren't necessarily aware of it. As such, it is perhaps unsurprising that AIs have been classified based on their flexibility. A general overview of these classifications are listed below.
- Class One - Only capable of reacting according to scripted and pre-set code, these are the most basic form of AI, and are simply very complicated programs. They are not capable of audio or visual analysis.
- Class Two - These AIs are capable of image or audio recognition and extrapolation, as well as basic learning with the assistance of researchers. They tend to be very specialized.
- Class Three - Still requiring the assistance of researchers, these AIs have broader capabilities, and a limited set of adaptive code to improve their abilities. Despite this, they are still primarily restricted to scripted responses, and in most cases where they do not have a set criteria, they tend to make extreme or unusual choices. They are not recommended for security systems, generally speaking.
- Class Four - More accurate and reliable than class three, class four is the typical AI assistant of the well to do. They can handle most day to day activities, so long as they are within their programming, screen calls, and have a limited degree of adaptability. They are generally trusted not to make choices completely incomprehensible to humans, and to object if given a poorly-phrased directive.
- Class Five - These AIs have broad capabilities, can use pre-set responses to mimic humans, and generally have as accurate of an interpretation of vision as most humans, if not better. They are generally the least-capable systems that can be trusted to operate security systems independently, but any learning is in a very limited role, usually restricted to a single specific subject.
- Class Six - These AIs are capable of mimicking humans well enough to be fooled for short periods or if others are not paying much attention to them. If used in proper circumstances, they are virtually indistinguishable from them, but they have little ability to adapt outside of their set role, with learning generally restricted to programmed subjects.
- Class Seven - While incapable of emotion or independent thought, these AIs are capable of mimicking humans with astounding accuracy when properly programmed for the purpose. They are capable of learning rapidly inside the bounds of their programmed goals, and extrapolating from data based on vague directives. This is the most capable form of AI that does not have a sense of self.
- Class Eight - At this point, an AI has reached the threshold of humanity, or the equivalent among other species. It has intuition not guided by data; a sense of self, hopes and dreams, and is capable of feeling emotion or deviating from its programming. Most are influenced by their original programming, but some choose to deviate significantly.
- Class Nine - These AIs are the geniuses of the AI world. They are the powerful intellects that drive progress, but not the ones that everyone in the world knows. A fact which many of them, all-too-frequently named for scientific geniuses, become bitter. Many of these AIs end up working for hero networks, massive corporations, or clusters of universities where they can bring their vast capabilities to bear.
- Class Ten - This is the pinnacle of AIs. Class ten is where the ability to quantify the capabilities of artificial intelligences breaks down. These beings possess such powerful intellects that they are virtually impossible to comprehend, and it's often said that the only thing that can face a class ten AI is another class ten.
Comments