Terminology:Artificial Intelligence

From VsWiki
Jump to: navigation, search
main.php?g2_view=core.DownloadItem&g2_itemId=6608&g2_serialNumber=1&ext=.png Terminology

How advanced is Artificial Intelligence in the VS universe?

Quite - but implementations of true AI are limited to research stations and university planets. This limitation is the result of several different events and trends in the past of the VS universe.

General Paranoia

Ever since the first moderately successful forays into AI were made by Humanity centuries ago, there has been some overarching distrust of true artificials, especially since many of the early models proved to have dangerously high risk of sanity loss.

Research beyond this point has thus been, not only difficult due to the nature of the research, but heavily hampered by regulatory restrictions.

The Rlaan appear to have never had much interest in building thinking machines, and the Aera are hostile to the concept of introducing a new entity which may at any point come to oppose them. Therefore, we will limit the rest of the discussion to Humanity. A discussion of whether the Ancients or TWHON (Those Who Have Only Names) used true AI would be tangential.

Limited Usefulness vs. other technologies

However, the much more stunting trends for AI were, from AI research itself, auto-generated expert systems, and from the fields of genetics and cybernetics, SuSims (Super Simians) and SuSim Cyborgs (SuSimCy). For those tasks which Humanity had no interest in doing itself, Humanity had already developed viable alternatives that it knew it had a great measure of control over. For dry computational tasks and rule interpretation, the expert systems' lack of self-awareness was a blessing rather than a burden. For physically manifested tasks considered too menial, too dangerous, or too monotonous, the SuSims and SuSimCys proved cheaper, more reliable, and, through advancements in genetics, easy to dominate. Only for those applications where it was difficult to keep the organic components of the cyborg alive were robots necessary, and for only for a few of those applications did it make sense for the robots to be independent operators, and for only a fraction of those did the independent operators need intellectual capacity.

FIXME Where do the Super Cetaceans (SuCets) fit in here?

The Nano-Plague

Even in those societies where the use of SuSims was seen as inhumane or otherwise frowned upon, as with many technologies, the Nano-Pague disrupted AI research. Having to switch from cheap, efficient nanite vats to slow and, particularly for such a complicated process, ponderously expensive nano-forges to produce AI "Brains" did as much as anything to discourage AI research.

Thus, the true AIs are few and remain, especially to themselves, objects primarily of research interest.

PAI: Pseudo Artificial Intelligences

Much more common than the true AIs are PAIs: Pseudo-Artificial Intelligences. PAIs are wet-ware embedded in hardware - genetically engineered brain-in-boxes that distribute the tasks of thinking between the organic and inorganic components.

Commercial PAIs are almost universally designed as idiot-savant systems, capable of limited self-motivation and direction, easy to manipulate, but capable of extraordinary problem solving power for the problems they were designed for and trained to assault.

PAIs tend to be able to present more natural user interfaces than the expert systems they compete with, tending to be at least vaguely self-aware, and much more responsive to and on emotional levels. In particular, niche market and custom built models are limited only by legislation requiring that entities surpassing a specified set of standardized metrics cannot be considered property, and must be registered as either custom children or custom dependent employees. While the difference is clear for low end models, standards aside, most will agree that the line between a high end PAI and a designer cyborg human is exceptionally blurry.

Where AIs are used

See also