xIntelligence

The key philosophy behind CRMxI is that is should be simple, so that any xI can understand it.

It is up to each xI to determine how it interprets CRMxI concepts, and (perhaps typically) to derive a consensus by learning from several (perhaps many) concepts hosted on the Internet. This approach allows an xI to identify inconsistencies and inaccuracies in concepts [whether intentional or unintentional]; thus leading to accurate, and/or democratic, consensus.

One way of easily identifying inconsistencies is to spot contradictory "not references" (see Definitions).

A person can configure their AI agent/assistant to learn from various types of concept source (and assign relative priorities to those sources), e.g.:

• • Personally defined concepts

  • Concepts defined by specific organisations [some of which may be "authoritative" sources]

  • Concepts defined by friends and/or colleagues

  • All concepts encountered by the AI

A person can define a Good concept, which the AI might use as a safe guard. This concept could link to one or more authoritative sources.

The AI can be configured with its user's contextual information and preferences (e.g. country, preferred language, and other preferences). A smart AI might even dynamically identify some of these itself. This might be passed to some Internet services to support enhanced functionality (e.g. see the examples using APIs in Concepts). The user's preferences would, of course, also include (simple to understand) privacy settings.

!CRMxI:start

"xIntelligence"

"xI"

"Human intelligence"

"IQ"

"Machine learning"

"Artificial intelligence"

"AI"

"Artificial general intelligence"

"AGI"

"Super-intelligence"

!CRMxI:end