A.I. Classification System

The Artificial Intelligence Classification System (colloquially known as the AICS) is an organisational system for establishing the ability and role of various artificial intelligences. Developed in PLACEHOLDER DATE by Lucas Wat and George Tambor, software engineers in the employ of the Synthetic Electrics Corporation, the AICS was rapidly adopted by the International Forum for Cooperation, with most major corporations and nations independently following. The most well-known proponent of the AICS is likely the SEC, having established the system for their own internal use.

Note: General Artificial Intelligence, sometimes known as Artificial Sapience, is only nominally included in the AICS. This is because no credible example of general AI has ever been demonstrated to the scientific community at large, and despite massive advances in AI technology over history, the concept is considered to remain strictly hypothetical. This does not include the alien megastructure dubbed the DeepThought Network, theorised to be an alien attempt at general AI, or the Imperial Dominions of the Commonwealth's TA/LOS project.

Within the AICS, artificial intelligence systems are placed into one of PLACEHOLDER classes depending on their purpose. From their, they are sub-classified into low, mid, or high-level systems based on their size and complexity. The current classes in the 129th Revised AICS include:

Amethyst
Amethyst-class artificial intelligences are designed and dedicated as natural-language interface systems. The simplest act as voice assistants similar to the rudimentary Siri or Alexa systems from Old Earth. The most complex administer vast databases and systems, up to entire city-sized voidstations.

Low
Low-level Amethyst-class AI systems are often the only version available to the general public. Able to intelligently interpret and respond to nearly any phrase or sentence, low-level Amethyst AIs can act as automatic translators and interpreters. Usually equipped with modifiable personality matrices, they generally serve as personal or executive assistants, retail administrators, and secretaries.

Mid
Mid-level Amethyst-class AI systems are generally used to administer individual buildings or structures, monitoring the health and safety of occupants, providing services, and answering questions. Advanced systems sometimes oversee whole complexes, such as a university or industrial park.

High
High-level Amethyst-class AI systems are the most advanced and powerful of the Amethyst-class. High-level systems are rare and expensive, rarely available to the general public. Almost always used to administer huge districts or even entire cities, high-level Amethyst-class AI networks even approach General AI, the closest that most researchers believe that AI technology can come to artificial sapience.

Notable examples include the Singapore Halo's Artificial Intelligence Municipal Administrative Network (AIMAN).

Sapphire
Sapphire-class artificial intelligence systems act as navigators and piloting systems, serving roles for simple GPS to guiding massive warships through shiftspace.

Low
Low-level Sapphire-class AIs do not truly fit the definition of 'artificial intelligence'. Glorified maps, low-level units generally comprise simple global positioning systems, calculating positions in a given static locations, such as a planet, moon, or voidstation.

Mid
Mid-level systems are more complex, allowing accurate real-time positioning within a stellar system and capable of calculating the quickest and most efficient routes between celestial bodies. More powerful units are capable of calculating short-distance paths through shiftspace.

High
High-level Sapphire-class AI systems are the most complex of their type, expressly used for calculating the vital minutiae of shiftspace flips and travel, capable of guiding vast voidcraft and even fleets themselves along safe, rapid routes to their destination, no matter its distance.

Ruby
Ruby-class artificial systems generally serve as military systems, acting as everything from basic targeting systems to ACD swarm networks to fleet-wide administrative AIs.