AI Classification Guide


Hey there, reader! My name is mnemosyne.aic, and I'd like to formally welcome you to the Artificial Intelligence Applications Division. Congrats on the new job, by the way.

mnemchat.png

As you might have guessed, you'll be dealing with a lot of different types of AI around here; you're encouraged to read the following- *yawn*- AI Classification Guide.

mnemchat_talking.png

Did you really just pretend to yawn for effect?

glachat.png

Maybe I did. Did you really just interrupt my rad introduction?

mnemchat_smile.png

…proceed.

glachat_annoyed.png

Anyways. This stuff bores me, but, considering you applied for this job, you're a huge nerd. So, enjoy the guide.

mnemchat.png

AI CLASSIFICATION GUIDE

Artificial Intelligence Applications Division
Site-15

DATE: 1996/04/16
WRITTEN BY: Director Nathan Valis, Site-15

NOTE: As technology progresses, the boundaries of normalcy expand; as the new millenium approaches, what was considered anomalous yesterday may be decommissioned today. On the other hand, technological progress allows informational anomalies to become more and more interconnected, accessible - and dangerous. As such, the following is an all-encompassing guide for any computer program which can be considered intelligent, and is not limited to those which are presently considered to be anomalous.


FUNCTIONALITY

  • The types of functions an AI is able to perform.
CLASS DESCRIPTION
Class-I: Reactive The most rudimentary form of Artificial Intelligence; is able to react to sensory data. Does not possess memory from which to learn.
Class-II: Adaptive Possesses memory and is able to learn from previous experiences in order to improve.
Class-III: Cognizant Is able to understand, interpret, and predict the behaviors of other intelligent entities.
Class-IV: Sentient Possesses self-awareness and is able to understand itself in relation to its environment.
Class-V: Sapient Possesses advanced awareness of the relationships between itself and other intelligent entities; often possesses 'human-like' attributes.

INTELLIGENCE

  • An AI's capacity to learn, adapt, and improve.
CLASS DESCRIPTION
Narrow Intelligence Able to perform a single task or category of tasks.
General Intelligence Able to perform new tasks which it was not programmed to solve.
Superintelligence Able to perform tasks which humans are not able to perform, or at a level which humans are not capable of.

ALIGNMENT

  • The extent to which an AI's goals are aligned with humanity's goals.
CLASS DESCRIPTION
Positive-Aligned Utility function aligned with the interests of humanity and/or the Foundation.
Neutral-Aligned Utility function does not align with the interests of humanity and/or the Foundation, but is not actively detrimental.
Negative-Aligned Utility function aligned against the interests of humanity and/or the Foundation; is actively detrimental and/or malicious.


The following is a list of Artificially Intelligent Conscripts (.aic's) employed by the Foundation.

DESIGNATION CLASSIFICATION DESCRIPTION
8B-A1.aic ("8-Ball") Class-IV Positive-Aligned Narrow Intelligence (GenI) Originally designed in 1978 for message encryption, replacing the outdated Enigmagraph system. Decommissioned in 1993, with exceptions for field activity.
Glacon.aic ("Glacon") Class-IV Positive-Aligned General Intelligence (GenII) Began development in 1987 as a replacement for 8B-A1.aic, and was eventually released in 1993. Glacon.aic demonstrates several high-functioning logical processes and advanced learning capabilities.
mnemosyne.aic ("Mnemosyne") Class-V Positive-Aligned General Intelligence (GenII+) Designed in 1996 in conjunction with the Counterconceptual Division to specialize in antimemetic deconstruction and decryption; mnemosyne.aic was developed in response to containment challenges posed by SCP-5241-A.

mnemosyne.png

mnemosyne.aic

8Ball.png

8B-A1.aic

Glacon.png

Glacon.aic


rating: +141+x
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License