~/ w.wiki/ technology/

Artificial Intelligence

This is the general page for artificial intelligence, defined as any intelligence created by something other than natural processes, but by something that could be generally referred to as an intelligence itself.

Categorization

Artificial Intelligences are categorized in many different ways depending on the use case. There is no one correct way to categorize AI’s. Some are categorized in terms of functionality (Narrow/Weak, General/Strong, and Superintelligence), Learning Methods (Supervised, Unsupervised, Reinforced, Deep), Interaction (Reactive, Limited Memory, Theory of AI, Self Aware, etc.), or based on Autonomy (Assisted, Autonomous, Augmented), etc.

Watson Scale

(section derived from Watson Scale Page)

AI’s within this universe can generally be categorized according to the Watson Scale which is a logarithmic scale ranging from 9 to 0 (where 9 is least powerful, 0 is most), factoring in the intelligence, computing power, utility, etc. of an artificial intelligence. A standard Human would factor in as a Tier 6 on the Watson Scale for reference.

Also of note is the general distribution of AI’s according to their categorization. Like most things involving a distribution of some sort in nature, the distribution of AI’s according to their Watson Scale rating such as the power of earthquakes or distributions of wealth follows a power-law distribution

The distribution of AI intelligences according to their Watson Scale can be represented as:

\(P(x) = C(-x+9)^{-a}\) Where

  • P(x) is the probability that an AI being has a Watson Scale
  • C is the constant of proportionality used to normalize the distribution.
  • x is the Tier on the Watson Scale
  • a is the shape parameter which determines how steep the distribution.

Of note, the % of distribution roughly follows the Pareto Principle, in which 80% of the Wealth (or in this case, Power/Scale Quotient) is held by 20% of the population.

Visual Representation:

![[Pasted image 20240911203742.png]]

Types of AI:

AI’s, much like their categorizations, are hard to define by types. Nonetheless, some basic categorizations of types can be made. One thing of important note is that AI’s that fit best into this scale are ones that were invented or created by Humans or at least according to Human technologies or are one degree of separation from Human creation (i.e., an AI made by an AI, made by a Human). Indeed, most AI’s of all types are made in this manner, though there are some AI’s that were made without human intervention. These don’t tend to fit into these scales very well.

Programs (Viruses, Algorithms, Routines)

  • Software based simple programs, either continuous or terminating.
  • Runs extremely basic functions, sometimes as a constituent of a more advanced AI
  • Rogue Programs can exhibit lifelike qualities, such as spreading across the Open and Closed Net and modifying themselves to fill resource niches. Can also exhibit other lifelike interactions such as parasitism, and symbiotic behavior.
  • Fundamentally basic.
  • Watson Scale Distribution: Null

Machinas (Bots, Drones, Servitors, etc.)

  • Basic Narrow AI’s
  • Used in specific, utility tasks
  • Programmed for tasks
  • Watson Scale Distribution: 9 to 6 Use Cases:
  • Utility Roles
  • Basic Robot/Control Systems
  • Personal Ai’s and Kiosks

Minds (Cognitors, Neurals, Synths)

  • ‘Grown’ as opposed to created
  • Advantages of Abstract Organic Thinking with Computer-like Logic
  • Prone to Psychosis if overextended
  • Programmed as well as taught
  • Watson Scale Distribution: 8 to 3 Use Cases:
  • Complex System control (Buildings, ships, Stations)
  • Complex Adaptable Robots
  • Personas (Public Figures, Ambassadors)
  • Research and Strategy Roles
  • Frame Inhibitors/Net Watchdogs

Frames (Omni Minds, Archotechs, Archons, Cores)

  • Hyper advanced, incredibly smart AI’s
  • General Intelligence, exponential learning
  • Incredibly High Power Requirement
  • Hard to understand, control, and communicate with
  • Watson Scale Distribution: 2 to 0 Use Cases:
  • Nation/Continent-wide Management Systems
  • Military/Political Strategic Command
  • Scientific Research
  • Frame Inhibitors/Hunters

© mmxxvi npn

Click the Yin-Yang to toggle theme